It's such a simple idea that I don't know how it couldn't have possibly existed earlier. It doesn't require massive graphic capabilities, no CUDA NUMA GPGPGPU and the like -- it's an idea so profoundly simple that you'll be scraping your jaw off the floor with a spoon before you finish reading this blog post.

You run the 'game' and it sets up a directory structure to represent every possible state of the game. To play the game, you inspect the current folder, look at the currently available sub folders, and choose which one to navigate into. Each sub folder you see represents the next possible state of the game.

There is no executable program running when you are playing. The game is just your act of navigating the folder structure. It's a clever form of madness!

Here's some screenshots that show me playing the game, in a command prompt.

I've changed my prompt to just a 'greater than' sign, by typing 'prompt $g', to make it cleaner. Old-school DOS wizards do this kind of thing all the time. Hipster kids, try and keep up.

First I type 'dir /b' to get a clean view of the current folder.

There is only one folder, '__PLAY HANGMAN__' so I navigate into it by typing  'cd[TAB][Enter]'. The game has begun.

Looking in that folder I see the empty scaffold, three underscores (representing a 3 letter word) and the available letters listed down the left hand side. We're playing with a reduced alphabet to keep the number of permutations within a reasonable amount.

My first guess is the letter 'A'. To make this guess I type 'cd A[tab]'. When I type 'dir /b' to inspect the new state of the game I see that it was a good guess, as one of the letters of the word has been filled in. Also there is an 'x' next to the 'A' indicating that that letter has been guessed. That's all there is to it. Hear that pop? That was your cerebellum exploding through your temporal lobe.

If I try guessing 'A' again, all I find is a folder containing a file named 'You have already guessed that letter.' To undo my mistake I back out by typing 'cd ..' (You could, theoretically use this technique to undo any move, but that would be unsportsmanlike.)

A couple of moves later and I've guessed every letter:

The word is 'CAB'. There's no particular fanfare. Just the word 'WIN' declaring my victory.

Okay, now you've seen the whole thing I hope you're aching to not just download the game, but to write your own entry into this brave new genre. You could do for Gehtto-Folder-Games what John Carmack did for first person shooters!

I'd love to see an implementation of tic-tac-toe, aka, naughts and crosses. Also, hangman could be minified by using junctions (i.e. symbolic links) -- I'd love to see that implemented. The possibilities for new games are pretty much endless. Maybe six or seven.

Okay let me level with you. The mathematics of this idea were just ridiculous. For a full 26 character alphabet, I would've killed my little computer. (Care to accurately calculate how many folders are created for the given alphabet?) It's all based around N-factorial where N is the size of the alphabet. Factorial is not something you want to see in the real world.

I'm on holiday at the moment, so the only computer available to write it on was my dell mini, which doesn't have any serious coding environment. So rather than write it as a console app in visual studio (or even a powershell app) I wrote a javascript page that generates a batch file. It's very niche.

This also turned into an opportunity to learn the ins and outs of getting recursion wrong with javascript. The most common mistake you make with recursion in javascript, it turns out, is forgetting to declare a local variable. This means that the variable becomes global and your recursion goes nuts. Don't make that mistake. When your code is creating subfolders all over your hard drive it's a particularly painful mistake. It took me considerably less time to write the program than it did to debug it, and (considerably*considerably) less time to debug than the time to clean up all the folders left in unexpected places.

But the funnest bit was the ascii sprite code.

I had one array that showed the final hanged man:

var hangyPicture = [
'   _______   ',
'   I     I   ',
'   I     O   ',
'   I   --I-- ',
'   I    I I  ',
'   I         ',
'--===--      '];

(Side point... why wasn't my ascii art better than this? Because | and \ are not valid in folder names.)

And I had another array that showed how many misses were required before a given character of the hangyPicture was shown:

var hangyMask = [
'             ',
'             ',
'         1   ',
'       33233 ',
'        4 4  ',
'             ',
'             '];

Then, with those two arrays in place, I can work out the names of the folders to create. I walk through the mask, one character at a time, and compare the digit I find to the current number of missed guesses. If the digit I find is less than (or equal to) the current number of missed guesses, then I include that character in the folder name, otherwise I mask it out with a space. Ah, I give in -- it's easier to just show the code:

for (var h in hangyPicture) {
  // use hangyMask[h] and numberOfMisses to mask chars out of hangyPictures[h] 
  for(var i = 0; i < hangyMask[h].length; i++) {
    if (hangyMask[h][i] == ' ' || hangyMask[h][i] <= numberOfMisses) {
      folders[h] += hangyPicture[h][i];
    } else {
      folders[h] += ' ';
    }
  }
}

Okay. Now you've been enlightened with the future of ultra-ghetto folder-based gaming, go ahead and make your own. I can wait.

Here's the code

The accompanying program is a piece of javascript that creates a windows batch file. You run the batch file.

The only things that will survive the forthcoming nuclear+zombie apocalypse are cockroaches, javascript and CSV.

So I've written an open source javascript library, agnes.js, that the cockroaches can use for dealing with CSV from javascript. And in the brief pre-apocalypse era you can use it too.

Download the zip

I've written before about how CSV starts off looking easy, but quickly descends into a world of insanity. So in javascript land, let agnes handle all the nasty quirks and edge cases (embedded delimiters, qualifiers, nulls and so on).

I wrote a bunch of unit tests to go with it, so you can tell exactly what it does with each weird bit of input you can throw at it.

You can download it from agnes.codeplex.com

This was a particularly fun little bus project, and what I like best about agnes is the content you get when you download her.

The readme file has executable examples, that work by having a chunk of code displayed in a div, which is the exact code that is grabbed and executed when you click 'Try it'.

Browse source code

I like the unit tests that come with it, and the little unit test runner. I could've gone really overboard with it, but I stopped myself before it went too far.

And I like the sample that performs Csv to Json conversion, back and forth, back and forth. My favourite part of that is using html entities in the button titles, for left and right arrows.

Try it out

Now that I've got agnes out of the way, hopefully I can focus on my new bilion dollar idea, youmustget.com

Yesterday Joel told me, 'hey you really should read "The Big Short" by Michael Lewis' -- and I thought, yes, I'll write that in my list of books I must get and when I have a chance I'll order it and all the other books in my list of books i must get.

Looking through my iphone I was shocked to find I don't even *have* a list of books I must get. I've got a list of which Woody Allen movies I haven't seen, and I have a list of which Jeeves and Wooster books I own (or own twice) so I don't accidentally buy them again (or again again), but no "Books I must get".

And I realised it's not just books that I keep forgetting. It's everything in our stupid materialist world. It's gadgets, music, movies, games, t-shirts, toys, gifts for my wife, baby stuff, apps, power tools -- a whole big fat materialist world full of stuff I must get, that, realistically i'll never get, but which my inner taxonomist wouldn't mind having a list of somewhere.

Suddenly, before I could stop them, my hands had rushed off and spent 9.95 to buy all of the following websites...

• http://things.youmustget.com
• http://books.youmustget.com
• http://movies.youmustget.com
• http://games.youmustget.com
• http://gadgets.youmustget.com
• http://apps.youmustget.com
• http://stocks.youmustget.com
• http://software.youmustget.com
• http://dvds.youmustget.com
• http://jokes.youmustget.com
• http://vaccinations.youmustget.com
• http://tattoos.youmustget.com
• http://cars.youmustget.com
• http://piercings.youmustget.com
• http://hookers.youmustget.com
• http://starewarsstuff.youmustget.com
• http://drugs.youmustget.com
• http://websites.youmustget.com
• http://cakes.youmustget.com
• http://clothes.youmustget.com
• http://bikes.youmustget.com
• http://girlfriends.youmustget.com
• http://babystuff.youmustget.com
• http://diseases.youmustget.com
• http://travel.youmustget.com
• http://tshirts.youmustget.com
• http://shirts.youmustget.com
• http://wines.youmustget.com
• http://beers.youmustget.com

...and so on, because i bought 'youmustget.com'.

And now I have a slight dilemma. I already own too many urls that I'm not using (this brings my total internet empire to 11.5 sites) And I don't have time to execute on even the simplest ideas in my idea log.

So: how can I crank out a social recommendation/shopping app in under 5 bus trips!?

I want something that produces an output a little bit like this is why i'm broke dot com but where the recommendations come from people i respect, and with affiliate dollars flowing into my wallet on every recommended purchase. (And i'd like to come up with a less-sleazy way to monetize, if possible, but I do have to recoup that $9.95 somehow)

I've begun by creating a Trello board to capture all my ideas. I've got details for the home page, a products page, a user page, and a topic page. I've got a backlog of features that I won't implement at first, and I've got a list of ideas for what technology I'll use where. Next I'll put together some screen mockups and put them in front of my extended team of insult generators lunch buddies.

In the mean time I'll need to keep a cricket bat handy with which to repel the hordes of nodding VCs.

Thoughts?

When I finally starting shipping, Joe Cooney said:

I tip my hat to you - I think in the time I've re-written a few bits of UI you shipped a product.

...in reference to his awesome LogEnvy tool. But it's not a fair comparison, for two reasons.

Firstly, LogEnvy is a much more sophisticated product than NimbleText. LogEnvy does all kinds of powerful manipulation of events logs in a wide range of formats and intersperses it with all the kind of dazzling UI you expect from a WPF ninja. But most of all his statement is incorrect because NimbleText was almost six years in the making.

I wrote the first version in February 2005 and published it as part of a blog post soon after. I wrote it the very same day I came up with the idea for TimeSnapper. It was a stressful week when I desperately needed both of those tools.

Although I was already a .net programmer, I wrote the original code in server-side VBScript for easier deployment. It was just a dozen or so lines. I called it variously 'the data pattern merge tool', 'the merger,' 'the programmer's mate' or the 'text multiplier.' Starting from that day, and continuing right up until the present, I have depended on this tool absolutely daily.

I released a second version in July 2005, and called it 'the world's second simplest code generator'.

The source code is largely a comment block listing all the extra features I wished it had. (And almost all of which, by the way, are implemented in more recent verisons)

A complete rewrite occurred in May 2006 when I started huffing the XSLT pixie dust and thought XSL could solve everything. Its dynamic nature gave it big advantages over the vbscript implementation.

But it soon turned into an exercise in writing a fully fledged compiler and I knew I was headed in the wrong direction.

I wrote an "offline version" around that time, in Windows Forms (VB.net version 1.1) I was never happy with it and didn't release it. If I have the code for it, I don't know where. My clearest memory of it was the way parts of the UI would animate (using a Timer control, of course :-) ).

I re-wrote the tool in August 2007, this time in javascript. Why did I rewrite it in Javasript? Because I was momentarily trapped on a machine that had no development tools, but I still felt like cutting some code. I assumed that browser performance would kill the tool, but I was dead wrong. Javascript was the perfect language for the problem, and the performance of Javascript was just beginning to take off as the browser wars re-ignited.

It was a clear victory for Atwood's Law:

"any application that can be written in JavaScript, will eventually be written in JavaScript."

Or a pre-cursor to Bambrick's Premonition:

"I don't know when skynet will take over. But I know what it will be written in: JavaScript."

I wrote a (playful) WPF version in Oct 2007 (in C#).

I wasn't happy with the tooling for WPF (maybe someone should fix that?), but more-so the dynamic behaviour of javascript trumped C# for this tool, so the WPF version never went anywhere.

Around that time I wrote other undocumented, unreleased versions in powershell and one in ruby (shoes). Powershell was well suited as a language, but lacked the reach of javascript, plus the performance (at that time) was dreadful. The shoes version dissappeared around the same time as _why. (Coincidence? I think not.)

Late last year I decided to release it as a windows app, using a polished and thoroughly bus-tested version of the javascript implementation, wrapped in a browser control inside a skinny winforms shim. Early this year I went live with it, under the new official title of NimbleText. I've hunted down bugs and added features in the months since.

The latest release looks a little more metro than any previous version, as I stripped away a lot of the padding and extraneous lines, simplifying the appearance (and improving a few features as well, of course).

It's kind of amusing that after I chose to shun all of Microsoft's native application development toolkits (wpf, silverlight, and winforms) going with HTML for the UI, it seems like Microsoft have now decided to meet me half way, making HTML a first class citizen in Metro.

Anytime I'm faced with a new platform, I usually implement a version of nimbleText as my own personal hello world, and it looks like Metro will make it easy for me. Porting the heart of this baby to Metro should be a breeze. I'll still have to put in some design effort around how the supporting features are presented (what will happen to the menus, toolbars etc?). But I expect it will be a good ice-breaker.

Meanwhile -- if you ever write code or munge data, save yourself some time and use NimbleText. Both the download and the online version are FREE. There's no time-limit on the application and you only choose to pay if you want the extended features (customizability and commandline integration).

If you have features you want, please leave a comment here or send an email to supportexclude this bit@exclude this bitNimbleText.com

I think i get it now. In Windows 8 you have two modes: metro and classic.

Metro is for exciting responsive new apps that you can use while sitting on a designer couch.

Classic is the Windows of spreadsheets and complex applications that we've used in our day jobs for years.

It's basically a mullet. Business in the front, party in the back.

I'm looking forward to joining the Metro party. I don't know if the business guy is going to host a very good party. But his client list is so long that a lot of great people are bound to be there.

Maybe I'm a little worried it will be like the party that Rick Moranis's character, Louis, is hosting in ghostbusters:

Louis: I'm givin' this whole thing as a promotional expense, that's why I invited clients instead of friends. You havin' a good time, Mark?
[heads across the room, greeting other guests]

It's all going well until a terror dog turns up and terrifies the guests. But I digress.

At first I thought that this metro-fad would only have as much take-up as designing sidebar gadgets in Vista (i.e. approximately none). But the killer features to attract devs is the app store. Having your apps readily available to (potentially) hundreds of millions of people, is a very tempting proposition and bound to attract a lot of developers, possibly including me.

So, does anybody wanna play Parcheesi?

'Cosby' is a small site for generating striped backgrounds (or gradient edges). You can incorporate the code from Cosby into your own asp.net/mvc applications and create striped backgrounds directly from within your Css file.

Once you've got Cosby in place, you can start to magically conjur stripy images out of thin air. Just reference a Cosby image as if it already exists, and it will cause the relevant image to be generated spontaneously and stored on disk for eternity.

For example, if you write: "background-image: url(/Image/Stripes/V/20/FFF/1/000);" then you will get a background with vertical ('V') stripes, alternating between 20 pixels of white (FFF) and 1 pixel of black (000). A moment later you decide you want the black stripes a little wider. Just alter the css (change the 1 to a 2) and they will be. No need to jump back and forth between image editor, filesystem, IDE and browser.

Cosby includes a playground where you can experiment with creating your own backgrounds.

Play with cosby

Get the code

Visit the site

But Why!?

While I was procrastinating about working on my talk for WDCNZ I wanted to mess around with MVC's 'FileResult' which lets you return a file from an action. So I resurrected and mutated the idea behind a Gradient Maker I wrote SIX years ago, using asp.net 1.0, and ended up with this latest monstrosity. (Cosby also lets you generate gradients, though I haven't done much with that feature)

The code is published at cosby.secretGeek.net, so feel free to grab it, criticise it, and forget about it.

And of course the name is a reference to Bill Cosby's infamous Cosby sweaters, which get their own detailed write up on wikipedia. I was recently heartened to see 'Cosby sweater' amongst the text generated by the hipster ipsum random text generator.

That's all for now.

I am just returned from an excellent holiday in New Zealand, wherein I talked about MVC 3 at the first ever WDCNZ in Wellington.

On the day of the talk I was deathly ill, so it was a very much less animated talk than I wanted to give, but still enjoyable (for me, if not the audience). I missed out on seeing the rest of the conference, which is a real bugger.

In particular I wanted to see Paul Irish on Html5, but I did at least get to meet him for a quick chat the day before. Other speakers I met the day before were pretty interesting to talk to and I very much wanted to see the talks by Kirk Jackson and Jeffrey Wegesin, among others.

Here's my slides, using an updated version of Higgins. You'll need to unzip them, then you can see the readme file, the slides themselves (by viewing higgins.html) or look at the included code.

Slides and Code for 'Live Coding Asp.net MVC 3'

The talk covers Razor syntax, Display and Editor Templates, custom Code Templates and custom Project templates. Thankfully, there is no video.

Joseph Cooney built a special utility to help with the talk, called 'BigNote' -- available from bignote.codeplex.com.

This is a handy tool for presenters. Invoke it with a hotkey (ctrl-F1) and then it lets you write in giant letters on top of the desktop. Very slick.

I had a slight panic during the talk when I realised that Joe Cooney, creator of ComputerUnlocked.com could have used the opportunity to slip any kind of nasty prank into the tool, just to punk me in front of a room full of strangers. But luckily he slipped up and missed the opportunity.

One other thing I ought to mention: the Dell laptop I bought in anticipation of the presentation turned out to be a total dud. I was only rescued the day before leaving Australia when Paul Stovell managed to hook me up with an old laptop of his, which performed flawlessly. Thanks Paul!

Web Dev Conf New Zealand (WDCNZ), is in Wellington very soon (schedule on lanyrd), and I'll be talking there about all things MVC 3. This is keeping me awake nights, I can tell you.

I've applied to talk at the stack overflow dev days conference in Sydney in October. That would be an absolute blast. I rank my chances pretty low, but my excitement is high regardless.

Recently I talked at QMSDNUG, on 'MVC 3: Third Times a Charm', alongside Joe Cooney. It was a lot of fun, and we allegedly broke a record with our crowd size. It's basically a preview of what I'll cover in Wellington, though some material will be in or out as I revise.

(Higgins got another workout, and had some html5 goodness baked in)

I promised to put up the references from my talk. They're mostly about customizating your MVC 3 experience in Visual Studio. Here they are.

If you use MVC 3, what are your favourite bits? What would you cover in a talk? If you don't use it, why not? And what are the sort of things you'd like to see deomonstrated?

Cool looking 404 pages are the new hotness. Github has an amazing parallaxing 404 page that allegedly cost more than any other feature on their site.

For a lot of sites, the 404 page is the most visited page, so it's worth getting it right.

The website for my new product, NimbleText, uses asp.net mvc. A framework I really enjoy. The gu wrote it on a plane. Before takeoff.

One of the more voodoo aspects of getting NimbleText.com into production was setting up a succesful custom 404 page. Here's what I came up with: check it out.

Some of the articles out there that cover custom errors in asp.net MVC seemed to be a little bit confused about exactly what is going on, many are out of date or incomplete and some are downright misleading.

So once and for all I want to give a definitive guide to error handling in asp.net MVC.

Here we are. Just eleven simple steps to follow for amazing results.

First, map a catch-all route in global.asax, at the end of your other routes. E.g.

routes.MapRoute(
   "404-PageNotFound",
   "{*url}",
   new { controller = "StaticContent", action = "PageNotFound" }
);

Second, create an Error Controller, like this:

public class ErrorController : Controller
{
   [AcceptVerbs(HttpVerbs.Get)]
   public ViewResult Unknown()
   {
     Response.StatusCode = (int)HttpStatusCode.InternalServerError;
     return View("Unknown");
   }

   [AcceptVerbs(HttpVerbs.Get)]
   public ViewResult NotFound(string path)
   {
     Response.StatusCode = (int)HttpStatusCode.NotFound;
     return View("NotFound", path);
   }
}


Third and fourth -- create custom views to handle the Unknown and NotFound actions above.

Fifth, create a page called Custom404.htm and add it to the root of your application. Use it to display a helpful, edgy and hopefully cool message. But don't be too edgy.

Sixth, add this to web.config, inside the system.web node:

<customErrors mode="RemoteOnly" defaultRedirect="GenericErrorPage.htm">
   <error statusCode="404" redirect="/Custom404.htm" />
</customErrors>

Seventh, add a httpErrors element inside the system.webServer node:

<httpErrors>
   <error statusCode="404" subStatusCode="-1" path="/custom404.htm" responseMode="Redirect" />
</httpErrors>

Eighth, inside your existing controllers, if there is no data to serve (e.g. if someone asks for a UserId that doesn't exist) then use this snippet of code:

throw new HttpException(404);

Ninth, inside Global.asax, look for the RegisterGlobalFilters method (it will be called during Application_Start), and add another global filter, to handle any specific exceptions in your code, like this:

filters.Add(new HandleErrorAttribute
    {
        ExceptionType = typeof(FileNotFoundException),
        View = "Custom404", // Custom404.cshtml is a view in the Shared folder.
        Order = 2
    });

Tenth, add a view called Custom404.cshtml under views/shared.

And, eleventh, take care to handle the global application error event, capturing and logging any errors, before redirecting to the views you're after:

protected void Application_Error(object sender, EventArgs e)
{
    Exception exception = Server.GetLastError();
    Response.Clear();
    HttpException httpException = exception as HttpException;
    if (httpException != null)
    {
        string action;
        switch (httpException.GetHttpCode())
        {
            case 404:
                // page not found
                action = "HttpError404";
                break;
            case 500:
                // server error
                action = "Internal500";
                break;
        default:
            action = "General";
            break;
        }
        // clear error on server
        Server.ClearError();
        Response.Redirect(String.Format("~/Error/{0}/?message={1}", action, exception.Message));
    }
}


Good.

Now your website is well and truly borked.

Every request will bounce around your application like a demented pinball tripping on acid.

When an error happens, a bunch of different code modules will go to war.

The victorious code module will tear out the entrails of all those who oppose it, and throw them in the visitor's face.

You'll get a different error handler depending on what version of IIS you have, what version of MVC you're using, whether you've deployed in debug or release, whether you're visiting locally or remotely, whether it's sunny or raining.

Your server will be reduced to a pile of smouldering rubble from which reboot is impossible.

But if you want to get truly weird, try setting up a custom 403 page. That's about three times as odd.

Or, switch to Dos on Dope*. Life will be simple again.

Sorry I didn't really have time to write a definitive guide to this stuff. Error handling in MVC is crazy man, I'm busy fighting a clandestine war against hackers and crackers and bots and beuracrats: I can't be documenting that kind of mystery wrapped in an enigma.

If someone does write a complete guide to error handling in MVC, that includes the meaning and interaction behind each of these snippets, plus how the interplay between IIS, asp.net and MVC truly works, then I will happily update this article to include a link to it, right up top.

* Don't switch to Dos On Dope.

My relentless campaign of jokes and nonsense will not be stopped.

And now, just a couple of long weeks later, here I am, happy to report I'm back in control of secretGeek.net.

Right when I was ready to migrate over to leonbambrick.com, I got an email from the Russian registrar, Regtime Ltd, saying:

Sorry  for answer delay. Domain was transferred onto you account.

The number one thing, I think, that helped get the site back was when a good friend, Madina, translated a lengthy email into fluent Russian for me to send to the Russian Registrar.

She re-structured the email to put the sob-story up front, all about how much personal meaning this site has for me, and the positive effects it has had on my life. I think that did the trick.

So what did we learn?

I learnt that passwords at google can be brute forced, if pop is enabled. This can be sped up by use of multiple IP addresses, or a botnet.

That's the most likely way they got access to my account. My password was 'good' by gmail standards but is now 'freaking solid' by any standard.

And I've turned on 2-step verification, plus all the other recommendations from part 1.

Thanks for the encouragement and support. It was dark times, but now the nonsense can continue.

If you have any information that could help me get control of my domain again please leave a comment, or tweet me (@secretgeek), or get in touch via my (now re-secured) email address, leonbambrick@gmail.com

From Russia with Love

On Monday 9th May, I checked my gmail account at around 3:40 in the afternoon, and I was confronted with a dark red message at the top of the screen (in the area where you normally see messages like 'Your email has been sent'). The message said:

Warning: we believe your account was recently accessed from: Russia. Show details and preferences | Ignore

I clicked on 'show details and preferences', and a new window opened with this message:

(click to see the message in context -- [note that my own ip addresses are redacted])

This was definitely not me. At 4:19 AM, and 5:40am I had been far too busy being fast asleep, preparing for a big week, to get to Russia and back for some casual email reading. So someone had infiltrated my email. The freaking out sensation began immediately. I couldn't move. I was frozen completely still.

I followed google's advice and immediately changed my password, then notified my wife. My mind was racing as to what the implications could be.

A little voice told me to check the trash. I was really hesitant, I think I knew the trash would contain something I didn't want to see.

In the trash I found two messages, received at 5:45am.

The content of the messages was to explain that my domain, secretGeek.net, had been succesfully transferred from my registrar, GoDaddy, to a Russian registrar, Regtime Ltd.

I felt completely unreal, like this was a dream, or a prank. I logged into my GoDaddy account to check on my domains. They all still appeared to be there. I took a few deep breaths. I felt it must be all just a daydream. I read the list carefully. No -- secretGeek.net was missing from the list. All of the others were still there.

I started sending emails to GoDaddy. I replied to the emails I'd found in the trash, telling them that it was a mistake, the result of my email being violated.

I lodged support requests with GoDaddy. When I went back to the trash to read the deleted emails again I found there were new messages in the trash. Emails from godaddy in response to my support requests had been automatically deleted. This was *weird*. It was a few minutes before the penny dropped.

I looked in my gmail filters and there were two new ones. The two new filters were designed to delete any messages containing the words 'GoDaddy' or 'Webnames'. (I later found that 'Webnames.ru' is the trading name of 'Regtime Ltd', the Russian registrar holding secretGeek.net)

This was starting to become very 'real'. This wasn't some kind of misunderstanding. It was definitely a thing. It was a deliberate, targeted attack, by a person or people who knew what they were doing and had either done this thing before or had thought it through very carefully.

I turned to my friend, Joe Cooney, to show him what I'd found. He was equally gobsmacked. He told me to call GoDaddy on their support number and see what they said.

A whois request on secretGeek.net showed that my name servers were still listed, but the registrar was Regtime ltd, and the contact information was all blank.

The support guy at GoDaddy told me to contact the gaining registrar and find out what their policy is for transfer disputes. He told me there was nothing GoDaddy could do (turns out this was bad advice, by the way)

I was busy worrying about what else was in danger. My wife was already on the task of changing every password, keycode, and security question associated with all of our bank accounts, credit cards, online identities, home pcs etc, and checking at each of these places for any trace of unusual activity.

In order to send a support request to the Russian registrar, I had to sign up as a member (and I noted with some distaste that they sent me my new password in plain text.) I sent them multiple emails explaining what happened. Then I re-sent each of those messages, in Russian this time, translated by translate.google, and including the original text.

I worked late that day to make up for the lost time and somehow managed another check-in before I went home.

That night I called GoDaddy again, hoping to get more help. By now I'd read about the similar case of David Airey where a gmail vulnerability led to a domain hijacking, which also involved godaddy. This time I got a more helpful support staff member. She told me to fill out the 'dispute on transfer away' form at GoDaddy, which I promptly did. This meant that the right people at godaddy would be notified, and they'd contact the Russian registrar to request the domain be returned.

Wrapping up part 1.

While there's more that's happened since, things haven't really progressed since that day at all.

I still haven't regained control of the domain. I'm waiting on word from the Russian registrar. I remain hopeful that any minute now I'll get an email from them saying "Ah, we're sending back your domain!" -- but my hope is fading.

The name servers haven't been changed, so my content continues to be served up as always. If it does go down, the plan is to launch a new and improved secretGeek at LeonBambrick.com. Announcements or news will be via twitter (@secretGeek) or from my business website, NimbleThing.net.

If you have any information that could help me get control of my domain again please leave a comment, or tweet me (@secretgeek), or get in touch via my (now re-secured) email address, leonbambrick@gmail.com

In the next part of the story I want to discuss gmail security, all of the possible ways I could've been hacked, and the one I think is most likely. But I can't wait until then to put out the following checklists. These are steps that you can do to protect your email account and to protect your domains:

What can you do to stop this happening to you?

Protect your email via:

• Change your password often. For example, right now.
• Make your password very strong.
• In gmail: disable POP.
• Ensure your password recovery options are as strong as your password.
• Make sure 'always use https' is selected.
• In gmail: check for unusual activity (it's available at the foot of the screen).
• In gmail: don't let any other apps be associated with your account.
• Use two-step authentication.

Protect your domains via:

• Pay for private domain registration.
• Pay for enhanced transfer protection. (e.g. at GoDaddy they call it 'Deadbolt' protection, and it means they require you to send them identification before they'll approve a transfer.)

A person with a Russian IP address broke into my email account and used that access to transer ownership of the secretGeek.net domain to a Russian registrar, webNames.ru.

The full story is jam packed with intrigue, daring adventure, black ops helicopters, sexy double agents, Ukrainian php developers, and a liberal dash of nerdy security best practices.

I will tell the full story once I have secured my domain again.

In the meantime, if your beloved secretGeek.net is suddenly replaced by a spammy porn site or variant thereof, do not despair. Follow me on twitter (@secretGeek) or head to my business website (NimbleThing.net) where (if necessary) I will mirror the original secretGeek content and continue blogging until I've taken charge of the domain again.

So here was the problem...

On one of my machines, in recent months, anytime I tried to run an asp.net application that attempts to connect to a SQL Server database I receive this error:

Login failed. The login is from an untrusted domain and cannot be used with Windows authentication.

This is in a scenario where everything is local: I'm on a home computer, using a local database server. And it doesn't matter if I'm using IIS, webdevserver, or IIS Express. It doesn't matter if I'm trying to connect to SQL Express or SQL Server 2008 r2. I can enter the credentials using the machine name or IP address. The result is the same.

Looking in the Windows Event Viewer I see:

SSPI handshake failed with error code 0x8009030c, state 14 while establishing a connection with integrated security; the connection has been closed. Reason: AcceptSecurityContext failed. The Windows error code indicates the cause of failure. [CLIENT: ].

Most of the 'solutions' online were not applicable because they involve domain issues. This is just a local machine.

One particular forum message involved a guy with a similar setup to me. He solved the problem for himself, but left behind only this very abbreviated and somewhat cryptic explanation:

Solved. Traced the prob. to my net setup. The clue was this a msg in the Win System Log from Lsasrv about the target name and it showed the fqdn. Did some checks and the fqdn didn't look right, cleaned up my network config (I had customized it for another project), and now it's workin.

The 'fqdn' is the fully qualified domain name. I found three different ways to find my fqdn, and one of them disagreed with the others! So this was a smoking gun.

Method 1 for finding the fqdn: When I right clicked on "computer" in the start menu and looked at the properties, the 'full computer name' was reported as 'Leon_xps'.

Method 2 for finding the fqdn: When I ran "ipconfig /all" and got these values for Host Name and Primary Dns Suffix:

   Host Name . . . . . . . . . . . . : Leon_xps
   Primary Dns Suffix  . . . . . . . : 

(Add them together and you get a fqdn of just 'Leon_xps'

Method 3 for finding the fqdn: In cmd when I ran 'ping -a Leon_xps' I saw a very unexpected result!

It said:

Pinging leonx_xps [192.168.1.2] with 32 bytes of data:
Request timed out.
Request timed out.
Request timed out.
Request timed out.

Notice that it's changed the name 'leon_xps' into 'leonx_xps' and come up with the ipaddress of '192.168.1.2'. instead of what I expected (such as a loopback address, 127.0.0.1... or some IPv6 nonsense)

There's only one kind of alchemy I know of for turning a good name into a crazy ipaddress: and that's the magic of the HOSTS file.

So I looked in there for 'leonx_xps' and I found it. Found this crazy nonsense line, probably added by myself during some other wild-goose-chase for a solution to some other messed up problem:

192.168.1.2 leonx_xps leon_xps

Once I'd commented out that rule, the asp.net application immediately started to connect succesfully to SQL Server 2008 r2. And I can move forward with my testing of Massive and Dapper.

But just mentioning the probable cause of this (not cleaning up during an earlier problem-panic) I remember that earlier tonight, while frantically trying to fix this problem I added some stupid rules to the windows firewall. I had better go and remove them now before my computer gets owned. Hey wait a moment!? Why is my CPU at 100%? Be right back...

I gave a talk at Brisbane's alt.net recently. You can download the slides here (1.5M zipped) or:

View the 97 slides online here

                (use arrow keys to advance)

The theme was 'Coding on the Bus' -- the optimisations and approach I've used for cranking out a dozen mini projects on my daily commute in the last year.

A bus is a seemingly counter-productive environment. As I say in one of the slides:

If your workplace was like a bus: everyone would quit.

And yet I'm producing a lot of stuff while on that short ride, definitely pushing the limits of my abilities. So this talk was an attempt to extract all the lessons I could from this real life productivity success story.

TL;DR;

Here's the quick version:

#1 Internet is bad
#2 Keep it damn simple
#3 Value your tools/libraries/snippets
#4 Don't be a hero

The people at Brisbane Alt.Net are a very cluey bunch so it was great to get their input on the whole thing. They were a very attentive and helpful bunch. It was a massive buzz talking to them. I highly recommend it.

...And Introducing Higgins

Naturally, I wrote the slides about writing on the bus, while on the bus. I don't have Powerpoint on my bus netbook, so I also had to write my own html-based slide presenter, which I've codenamed Higgins.

I've since added higgins to codeplex. You can get Higgins from codeplex and use it yourself, if you're so inclined.

Visit higgins at codeplex

Download higgins here

"Test-driven Dentistry (TDD!) - Not everything should be test driven"
- Michael Fogus

I'm no TDD-lover, not by a long shot, but I dismiss the claim that you shouldn't do test driven dentistry.

I think you probably *should* do test driven dentistry.

First write your test:

Setup:
   Put on lead coat.
   Perform X Ray.
   
Test:
   Assert: There are no unfilled cavities.

Teardown:
   Remove lead coat.

And then:

Run the test: it fails.

Fill the cavity.

Run the test: it passes.

The only special nuance of test-driven-dentistry is that you must resist the urge to refactor the mouth, as this would require re-executing your tests, which may cause unnecessary irradiation of the patient. Side effects are an encumbrance but not a show stopper.

I put out a minor update to NimbleText the other day and called it the 'less crashy' edition.

Here's the backstory.

I've received intermittent reports for a while that NimbleText would crash on startup. But I couldn't reproduce the problem, and even the people reporting the problem could rarely reproduce the problem.

After I put out a major update on Mar 17 I got this crash report from Scott Hanselman, via twitter:

@secretgeek I was going to tweet about it, but it crashes on startup for me.

Now this is vexing. I was on a bus at the time, with limited resources. I asked for a stack trace. (And I asked nicely, not like this guy)

But unfortunately:

@secretGeek can't...doesn't even start. Watson's before the UI shows up.

...which is even more vexing. (For the younger readers: to 'Watson' is to crash, and refers to Dr Watson. Internally at microsoft, this nomenclature is still used.)

So I asked for system specifications from the crashee, and learnt:

@secretGeek win64, ie9. Visit site, download, run, crash

Note that twitter is sufficient for most of the pertinent information, provided the crashee in fizzbin savvy. Non-fizzbin customers often have to give you the full backstory of their life and times before confirming that they are indeed using a Windows computer.

Well I haven't used IE9 myself, so that's the 'odd man out' for me. In fact as a chrome user, and occasional firefox user, I haven't tried running it from IE at all. When you write a windows application you don't think that testing various browsers should be a part of your regime.

But I had an immediate counter claim from OJ:

@secretGeek @shanselman I have Win7 x64, IE9, and it works fine for me. Download, run, all good. Leon, are you using 3rd party native dlls?

NimbleText is a very simple application. It has *no* dlls. It does use obfuscation though, and this is where my suspicions immediately shifted (wrongly as it turns out).

Once I was at a real computer again I tried downloading the app from IE and running it. It worked fine for me. Perhaps I could re-badge it with the 'works on my machine' certification and be done with it? No, these gloomy ruminations aren't helpful.

I tried again and again, and the trick soon announced itself.

When you click on an exe, what do you do next? Do you choose "run" or "save"? Personally I always choose "save". This became ingrained behavior in me many years ago. So it never occurred to me to press 'run'. But consider the psychology of the crashee: Scott Hanselman has personally installed every single piece of software on the entire internet. To allow him to perform this incredible feat he is going to be taking shortcuts anywhere he can. So instead of downloading the file and then running it, I tried just pressing "run". Instantly: the app would hang, then crash. Bingo.

Once you've got a repro, everything is straightforward.

The line that failed was an Assembly.Load:

var a = Assembly.Load(assemblyShortName);

Why does this fail when I run it from IE, but not when I download it and run it?

I assumed that some kind of security limitation didn't allow an Assembly Load in that scenario.

Rather than solve the problem of why the Assembly.Load failed, i looked at the code in question and found I could instead just use:

var a = Assembly.GetExecutingAssembly();

This code does *not* fail when the app is executed directly in IE. So I updated NimbleText and put out the new, less crashy, release.

But the mystery still wasn't solved to my satisfaction. I wanted to learn exactly why the Assembly.Load would fail. If it was a security exception, why didn't it look like one? I googled to no avail. I binged without success.

Next I built a small console app that tried many variations around the problem. I found I could get an Assembly.Load to succeed if I used the full name of the assembly, like this:

var a = Assembly.Load("NimbleText, Version=1.2.0.29995, Culture=neutral, PublicKeyToken=null");

What's the difference? I was expecting to need to use FUSLOGVW.exe the assembly binding log viewer, something I don't look on with any great excitement.

Then did I noticed that when you run an app directly from IE it not only gets downloaded to temporary internet files, it also gets *renamed*. I tried this out locally and saw that renaming the file was the killer. Any old copy of NimbleText would crash if I renamed the executable.

The CLR assembly resolver expects that your assembly name is your assembly short name plus .dll or .exe. Hence, it's quite rude of IE to rename the file under you. So I blame IE. I'm not the first person to blame IE, and I won't be the last.

But I also blame myself, or rather my code. I'll no longer be using shortnames when trying to resolve an assembly. It's strong names or nothing as far as I'm concerned. As of now, I scoff at short names.

And, you should use NimbleText for all your text manipulating, data extracting, code generating needs. It boasts less crashiness.

Count the Arrows on the Toolbar

Toolbars are kind of ridiculous. You start off with row after row of cryptic icons which mean almost nothing to you. Even after years of using a program you've probably only ever used half of them, and have no clear idea what the other half do, and probably don't need them anyway. Not to mention that it's so hard to come up with meaningful icons that most of them end up just being an arrow of some kind.

To (supposedly) overcome these kinds of problems Microsoft invented the 'ribbon,' but it hasn't made its way into Visual Studio yet. I think a different solution could be better:

Toolbars could work the same way as the taskbar does in Windows 7:

It would start out empty.

Clean. Beautiful. This is far preferable to today's approach: starting out full of clutter.

Thereafter, when you run a command (by selecting it from a menu option, for example), it would show up on the toolbar (in the same way that running an application in Windows 7 show up on the taskbar.

After that you would be able to right click it and choose to 'pin it' in place, so it stays.

There would need to be some difference in the behaviour, since commands and tasks don't work the same way as each other. Tasks are long running things, so an unpinned icon (in Windows 7) stays visible only so long as the task stays running. Commands are short running things, so if we followd the model exactly they would flicker onto the toolbar and then almost immediately dissappear before we have a chance to pin them. So instead the toolbar could show a 'rolling history' of (say) the ten or twenty most recently used commands. Or perhaps it would show all the commands used during this session. Or use a simple forumla to show commands that are used recently and or most often (similar to the way voting works at sites like reddit).

And you would be able to simply drag to rearrange the items on the toolbar, just like you can on the windows 7 taskbar.

I think this is better than a ribbon.

Some refinements of this idea are:

1. You might have a setting to say "Don't add a command to the toolbar if i've invoked it via a keyboard shortcut." Since you know the shortcut, you're unlikely to need the icon.

2. Some icons should only appear when you're editing certain types of files. The 'jumplist' for a given icon could let you quickly apply this kind of conditional showing/hiding.

3. Your toolbar settings could be exported, synchronised across computers (via dropbox) or shared with friends. They would be human readable.

4. It would be crucial to have an effective way to search for commands. This is true of the ribbon as well. (this part was well covered by Jeff Atwood in 2007).

See also: Rico Mariani discussing the future of Visual Studio (from 2008). He's no ribbon fan.

And a beta-tester of my blog tells me Hanselman has been covering similar territory in the last few days. That VS UI has just got to lighten up!

The History of every major Galactic Civilization tends to pass through three distinct and recognizable phases, those of Survival, Inquiry and Sophistication, otherwise known as the How, Why and Where phases.

For instance, the first phase is characterized by the question How can we eat? the second by the question Why do we eat? and the third by the question Where shall we have lunch?

In accordance with that philosophy, I've studied the complex decision making usually required to reach group consensus on a lunch venue, and embodied the resulting deterministic algorithm in this simple web site:

http://secretgeek.net/lunch/

So any day you're wondering where I'm gonna be for lunch, just visit that site and it should be correct. (For the next two weeks.)

I've recently gone through the process of getting email in place to support NimbleText, so I thought I'd document it while it's still fresh in the old mind.

The first thing you need to work out is how you intend to use email in your microIsv, and from this you'll determine what accounts you need.

For example, here's where I'm using it, or planning to use it:

1. When licenses go out
2. When support requests come in
3. potentially for 'mass mailouts'

I ruled out 'mass mailouts' but only for now.

To handle these simple cases I could have a whole bunch of addresses such as 'licensing', 'sales', 'support', 'marketing', 'leon', 'ceo' and so on. But that's exactly the sort of over-engineering you need to avoid. In a microIsv you need to be thoroughly minimalist in everything you do.

The simplest thing that can possibly work is to have exactly one account, 'support', so that's what I've done with NimbleText.

Step 1: Create the account.

I'm hosting my sites at Arvixe these days, so in their control panel I set up a new email account.

I also tell it to forward to my gmail account (this won't be enough to get the forwarding working, we need to tell gmail as well)

I also told it to act as a catch-all address. So if someone writes to 'unknown@...' etc then I should still get it.

Step 2: Set up 'send mail as' at gmail.

In gmail you need to configure another account you can send email from.

I seem to 5 or 6 of these in use by now.

Within gmail, you go into settings in the top right hand corner. Choose the tab called called "Accounts and import". In the 'Send mail as' section, click the button to "Send mail from another address."

Enter sensible values, and pick the defaults. To complete this process, gmail sends an email to your new account, that contains a confirmation code you need to enter back in gmail.

Step 3: Create a filter at gmail.

You want any new support requests, etc, to be highlighted, and grouped together.

So create a filter in gmail that applies a new label, named after your microIsv. Tip: for any label I really care about, such as this one, I pre-pend the name with an underscore so it is alphabetically near the top of my labels. (I have a lot of labels). In this case the label is "_NimbleText".

I also tell the filter to apply a star to all emails from my support address. I can leave the star in place until a support request has been followed up sufficiently. I treat anything with a 'star' as an open case.

(Yes, this is a poor man's bug tracking system. The startup package from fogBugz is free and powerful. I assume Atlassian's Jira has a similar deal.)

Step 4: Color the label.

Once the filter is created and the label is created we can then set a color for the filter, so that any new support messages stand out in the inbox.

Step 5: Listing email address online

Now you need to list the email address in a couple of places. I've put it on the about page, the help page and the purchase page.

I was *particularly* pleased with the idea of putting the email address on the purchase page, as I think it's just the kind of reassurance which might help people make the leap of faith required to put your money in the hands of a stranger.

Step 6: List email address in your application

Stupid me hasn't done this yet.

It's a good idea to make your email address available from the help menu of your application -- perhaps two links, one for 'add a suggestion ;-)' and one for 'report a report :-('.

Also, when the app crashes *hard*, your real clever applications catch the otherwise unhandled exception, and encourage the user to send them to the support address.

A more sophisticated approach is to send the crash reports over http. If you're integrated with Fogbugz, as we are with TimeSnapper, then you can use their 'bugzscout' dll to deliver crash reports, with customer permission. (also, here's an interesting codeproject article from 2004 on handling exceptions by some weird unknown guy called Wumpus1)

Step 7: When sending, pick correct 'from' address

This is pretty self-explanatory. If you're sending an email in your capacity as CEO of your new microIsv, pick the relevant email address in the 'from' drop-down list, rather than your regular email address.

Step 8: Sending email programmatically.

At times you'll need to send one or two emails programatically. For example when a purchase occurs and you want to send out a license.

You don't want those emails to be categorised as spam. There's a bunch of things to keep in mind, head to stackoverflow to read more.

When performing mass mailouts it becomes much more difficult to avoid being classified as spam (even if you're not sending spam) so when I've done that in the past I've outsourced it to a specialist company. I've used Campaign Monitor and was seriously impressed by the whole experience. I'd recommend you use them for any bulk promotional emails you need to send. (No, I'm not affiliated with them.)

Wrap it up already

That's it. That's every detail I can think of for getting email up and running to suit the tiniest of enterprises.

Comments, suggestions, feedback, bring it.

What is the NO Visual Studio Movement?

I don't talk about it much, but I'm the chief of an underground technical movement called the 'NO Visual Studio' movement, or NoVS to its friends.

NO in this case doesn't mean 'No.' We steal from the NoSQL movement and re-define NO to mean 'Not Only'.

In NoVS we respect that sometimes a set of lightweight tools can be more powerful (and certainly more enjoyable) than the 'one tool to rule them all' philosophy of VS. Sometimes low-power = high productivity.

The movement has its roots in this talk by Charles Petzold, 'Does Visual Studio Rot the Mind?' The final section, 'The Pure Pleasures of Pure Coding' is most relevant:

I decided to use plain old ANSI C, and to edit the source code in Notepad -- which has no IntelliSense and no sense of any other kind -- and to compile on the command line...

...there's no APIs, there's no classes, there's no properties, there's no forms, there's no controls, there's no event handlers, and there's definitely no Visual Studio.

It's just me and the code, and for awhile, I feel like a real programmer again.

He says everything right there in a nutshell. Every word is like strong whiskey.

In keeping with the NoVS movement, my favourite machine is a mini-laptop I use on the bus. It's a true 'No Visual Studio' machine where the only IDE is Notepad++.

Mostly I write javascript, and occasionally C# and Silverlight. NimbleText, for example, was developed almost entirely 'NoVS' style.

I've tweaked my setup lately so that I can compile entire .net projects from inside notepad++ -- and I thought I'd share that bit with you now.

Compiling .net projects in Notepad++

Download this file and put it in your Documents folder:

Download tryCompile.txt (rename it as .bat)

Rename it to tryCompile.bat... and always read batch files before running them on your own machine. (On a stranger's machine you can afford to be less cautious)

Once you've got that file in place, find notepad++'s shortcuts file. It should be here: %APPDATA%\NotePad++\shortcuts.xml

Edit that file. And under this tag <UserDefinedCommands> add the following node:

<Command name="MSBuild" Ctrl="yes" Alt="no" Shift="yes" Key="66">%USERPROFILE%\Documents\tryCompile.bat &quot;$(CURRENT_DIRECTORY)&quot;</Command>

Now, while editing any file that is part of an existing .net project, hit Ctrl-Shift-B. That will launch 'TryCompile.bat' which will in turn locate and build the correct project file for you. (It will also add an MSBuild option to the Run menu in Notepad++)

This works for C# projects, VB.net projects, Data dude projects -- anything that includes a proj file that MSBuild can consume.

If you have other tools, hacks or tweaks to further the cause don't hesitate to share them.

One of my buddies has turned into a bit of a Pinball Junkie lately.

That's a picture of him at right. He takes things to extremes.

His new goal in life is to get his name at the top of the high scores.

Which left a few of us wondering how data entry works on a pinball machine.

If there are only two buttons, left and right, it must be hard to enter your initials for the high score. My guess was that you'd scroll back and forth through the letters, but I wasn't sure how you'd confirm your letter, once you found it. Rhys was confident that you could do it with only two options, by some kind of binary tree navigation exercise.

How could a pinball machine inspire a programmer's editor?

The conversation quickly devolved into a design for an editor with just two buttons, one and zero. This would be a serious improvement over fuv.

You might have seen this famous cartoon before: 'Real Programmers Code In Binary'

(corrected -->)

(I'm not the only person annoyed by the extraneous use of an 'enter' key in that image... I've gone and fixed that.)

(And of course you've seen the infamous xkcd 378 which deals with the same topic.)

So Build It Already

As soon as I had a chance I started work on just such an editor, and it's available now from codeplex.

Visit ZeroOne on codeplex

And you can download it, too:

Download ZO.exe

Only two buttons are needed for input - Zero and One. So the editor is called ZeroOne, or 'zo' for short.

If it helps, you can think of one and zero as the left and right buttons on a pinball machine.

If you enter 'esc' (by typing 0011011) you enter 'Command mode'.

In command mode there are three commands - 'o' (open), 's' (save) and 'q' (quit).

For example, in command mode if you enter 'oc:\fred.txt' the first letter is the 'o' command and the remainder is the file path and name. So a file 'fred.txt' would be opened from 'c:\'.

To save a file you need to specify the filename. E.g. 'sc:\fred.txt'. Again, the first letter ('s') is the command.

To execute your command, enter a line feed character (0001010).

To exit command mode, enter 'esc' again (0011011).

If you press any key other than Zero or One, your document is wiped. (This is a fuvism) See screenshot at right.

If you send the wrong character into the editor, ZO can be quite forgiving: you can enter a Backspace by typing 0001000. There are no navigation commands though, so you should endeavor to enter the right text from the very start.

I've put together a short screencast to demonstrate ZO in action.

ZeroOne - the editor where you can only press Zero and One from secretGeek on Vimeo.

... and what was my friend's reaction when I showed him ZeroOne in action? I think his face says it all:

(If you don't have mercurial, download it and install it now, so you can follow along)

(This isn't a "joke" post by the way. Sometimes my sarcasm runs so deep that I need to stop and reassure people when I'm not being sarcastic. In particular, I need to reassure myself ;-) )

Say I'm creating an open source project, such as fuv. I create the project online, then clone it locally:

c:\projects>md fuv
c:\projects>cd fuv
c:\projects\fuv>hg clone https://hg01.codeplex.com/fuv

(If I'm just creating it locally, like if it's closed-source, then I begin instead with 'hg init.')

The very first thing I add is a .hgignore file. This is the file where you tell mercurial *not* to check in all of those pesky user-specific files that a visual studio project generates, and to avoid versioning the bin and obj folders etc.

I've stolen the .hgignore file from the Funnel Web project because 1) the guys working on that project are much more thorough and intelligent than me, and 2) it was the first .hgignore file I found lying around on my hard drive.

The .hgignore goes in the same folder as the .hg folder. It doesn't go *in* the .hg folder (you don't normally have to go in there at all). It goes in the parent of the .hg folder.

Now it's time to write the code. You could either copy code in from elsewhere (if you had already started it) or if starting from scratch, create a new solution in the folder you've just created.

Here's the bit that I think is clever/stupid. I put a 'post-build' action on the project, so that everytime I build succesfully I check in the work in progress.

It's as simple as adding this to the post build step:

hg addremove
hg commit

This is clever because it means that every tiny change you make is tracked. And it's stupid for the same reason. The main reason reason it's stupid is that it means everytime something has changed you are asked to enter a commit note in a notepad.

It's quite annoying at first. But soon you get into the spirit of it. You realise that it is making you act in a more mindful manner. You are concentrating on every build, more aware of what you are doing, and better focused on your tasks.

Still, that bit could be improved, and it definitely wouldn't be to everyone's taste.

When you've completed a feature you push your work to the server using hg push, like this:

hg push

You have to enter a username and password at that point. Or, if you're pushing often you can store the username/password in mercurial's configuration.

There's a lot of different places and ways you can store this info. I'm going to describe one in particular, but I'm not saying it's the best one. Opinions or improvements welcome.

Remember I said earlier that you don't need to go into .hg folder.... well... go in there. There's a file called 'hgrc.' Editing it using something better than notepad (as it seems to have unix-style line-endings rather than windows-style line-endings (aka '\n' not '\r\n', aka 'Lf' not 'CrLf' aka 'char 10' not 'char13 char10').

Add three lines to it, to tell it the user name and passowrd, and which site they refer to:

[auth]
codeplex.prefix = https://hg01.codeplex.com/fuv
codeplex.username = secretGeek
codeplex.password = joshua

(Joshua is not my real password by the way. And the word before the "." can be whatever word you want as long as it's consistent across all three lines. It's just for grouping purposes.)

Here's the stackoverflow question on mercurial passwords that I got this from.

Automate your pushes!

Once you've got your username/password in place, you can make the push a post build action too!

To make it so that a push only occurs on a 'release' build, try this:

if /I "$(ConfigurationName)" == "Release" hg push

After that you've got an automated, integrated approach to version control even for the most insignificant of little projects.

Getting it setup and in place takes just a minute or two for a new project.

Some of the tweaks I'd like to make to this system are:

1. use some kind of a custom script to generate a proposed commit message. Perhaps it could look in your todo list (such as nextaction) to grab the item you are currently working on.

2. somehow 'flattening' the local commits when pushing, so that all of the in-between personal revisions are squished out of the way.

Also, I've been thinking about using mercurial (with 'auto commit' as described above) for managing a todo.txt file. Every time you save it commits, and a script looks at the diffs to see what tasks were deleted, marked as done, edited, added, etc. While it's an intriguing idea, I think it would need you to be very disciplined and somewhat restricted in the way you use the todo list.

(nexaction, by the way, lets you run custom shell commands when you complete a task. So you could use this to commit the file every time you complete a task. It's a feature I've never discussed with humans, out loud.)

'fuv' is a programmer's editor that is an excellent replacement for vim.

If you've been using vim for a while, you're probably frustrated with its lack of hardcore power.

In vim it's far too easy to wimp out and use 'insert mode' to type text directly into a document.

fuv fixes that.

In fuv, there is no direct entry of text.

There is just a regular expression search/replace function.

*All* editing is done by searching and replacing over the existing text, using regular expressions.

And every search term *must* contain at least one non-escaped regular expression special character.

If, for example, you tried to replace "a" with "b" -- fuv wouldn't run it. It would emit the 'why bother?' error message and discard your document.

Instead you'd want to try replacing "[a]" with "b", only then will fuv consider your input worthy of applying.

If no matches are found for any given search/replace then fuv revises its opinion of you and discards your document. It may, on a whim, locate another document on your machine and discard that too.

fuv's native format for storing a document, a .fuv file, is not the resulting text but rather the series of replacements required to recreate that document from scratch. This is particularly helpful for avoiding merge conflicts from source control systems, mostly because it ensures that no one else will contribute.

Oh, and it's worth mentioning that fuv does not let you see the document itself. Upon launching fuv, all of your monitors go blank. Thereafter you must issue at least one command every minute or fuv assumes you have died and summarily discards all your work, overwriting every byte on disk at least 8 times to avoid recoverable data remanence. And it's certainly best to keep fuv busy, as it only uses *spare* cycles for searching your hard drives in order to locate configuration files from competing editors that it can delete, reverse, or otherwise insidiously corrupt. fuv is playful. i guess.

To exit fuv you need to enter the termination command ::q&*$&(#&. Be careful to get it right. And please, whatever you do, don't try to "paste" in the termination command, or any other command. fuv treats all use of the clipboard as a dire insult. I've promised not to reveal what happens, lest it should happen again to myself or the people I love (fuv is reading this as i type). Let me put it this way. Have you seen any of the "saw" films? fuv makes "saw" look like the dux of charm school.

Once you tire of fuv, it's time to consider emacs. Be careful though; emacs is no picnic. After you've been coddled by the niceties of fuv, emacs can seem outright cruel.

I've implemented the core features of fuv and put the source on codeplex. Download the source, compile it, or, if you're feeling particularly foolish, get the binary here.

Get "fuv.exe" Now (not a sane choice)

Visit fuv on codeplex (wiser)

Sorry 'Bob Vogt' but you've sent me the worst recruitment spam I've ever read.

I've taken the time to catalog precisely what it is that makes your recruitment spam email, out of thousands of other competing emails, so very definitely the worst I have seen, will ever see or could ever see.

Here are the 7 ways you fail, in order from least to most significant.

1. Bad: You've sent me unsolicited email.
2. Bad: You've included a massive image in your email.
3. Bad: The job is in the US. (I'm in Australia)
4. Worse: The job is in Dayton OHIO.
5. Worse: Your grammar is an atrocity.
6. Worse: The job involves sharepoint.
7. Worst: You're using comic fucking sans.

I spent about a fortnight in a private beta, receiving helpful feedback, bug reports and suggestions from Simon Harriyott, Atli Oddsson, Joseph Cooney and Rhys Parry.

A private beta is a great idea if you're trying to get your product up to the public stage. There's ample evidence that you only need a small number of active participants to run a successful beta. I kept the feedback from the participants separate (no shared email threads, for example) so that people's opinions wouldn't be unduly influenced. I quickly found the opinions overlapped on most of the mistakes I'd made. When the bug reports began to slow down and overlap, I figured it was time to move to public beta.

Meanwhile, one of the bugs discovered after launch was that I don't actually receive the payments that people make. They would end up at a kind of dead letter office at paypal. That was a pretty bad bug. I've fixed it now. Feel free to try it out ;-)

I just released my new product, NimbleText. Technically I guess it's in public beta now. This means you get the opportunity to be one of the cool kids performing beta testing.

NimbleText is a downloadable version of 'the world's simplest code generator.' It's free, and adds a few features to the existing online version. If you register it, it adds more features again.

Get "NimbleText.exe" Now

I'd appreciate any feedback you've got. Maybe you have:

• Suggestions for new features
• bug reports!
• feedback on the website (NimbleText)
• questions
• comforting words

Feel free to email me: LeonBambrickexclude this bit@exclude this bitgmail.com ]]> Fri, 28 Jan 2011 22:02:58 GMT Leon Bambrick http://www.secretGeek.net/nimbletext_is_alive.asp http://www.secretGeek.net/get_jonathan.asp Jonathan Coulton is a much loved musician amongst your nerd types, like me.

He's put out quite a few free songs in MP3 format over the years too, and these remain available online.

I found Rob Olson wrote a script with wget (and another with curl) that downloads all the free songs for you. So for my Microsoft-enslaved brethren I now provide a translation into powershell-ese.

Here it is -- Get-JoCo.ps1

Just download it and run it in powershell. (You may have to set the execution policy to gawker-level first. And don't read the script before you execute it or anything. Security is for girly-boys who don't have the guts to run my code at full trust.)

And of course, like all the repetitive code I write, this was written with NimbleText.

It's natural at this point to also encourage you to buy one of his albums, donate money to him, or perhaps whistle one of his tunes while performing a good deed that advances civilization. Your call.

Get "get-JoCo.ps1" Now

NimbleSet helps you compare two lists.

Go play with it now. If you can't figure out what it does, then come back here and I'll demonstrate it for you.

You're back? Okay, let's talk it through.

Say you have two sets of data, two lists, and you want to compare them. We can demonstrate this problem with the awesome power of a Venn diagram:

The venn diagram has three interesting areas:

• the bit on the left ('Left only')
• the bit in the middle (the 'intersection')
• the bit on the right ('Right only')

Imagine our two sets are: 'the beatles' and 'the gospels of the new testament'. (Sorry for my Anglo-Christian bias here, I couldn't find another well-known, short, pair of sets with an intersecting member)

If we want to see what items they have in common, click the 'intersection' button.

result (1 row)
john

To see which items are in the list on the left but not in the list on the right? Click 'left only.'

result (3 rows)
paul
george
ringo

To see the 'right only' items click 'right only.'

result (3 rows)
matthew
mark
luke

To grab all the items press 'union.' (You may notice that items from both lists will be repeated accordingly).

result (8 rows)
john
paul
george
ringo
matthew
mark
luke
john

To find the distinct members of the result set, press 'distinct' (under the 'result')

result (7 rows)
john
paul
george
ringo
matthew
mark
luke

To sort the results, as you often need to, press the sort button, under the result.

result (7 rows)
george
john
luke
mark
matthew
paul
ringo

There are also buttons to invert the list (in effect a sort descending), or to copy the result back up into either the left or right list.

That's about it.

It's much less powerful than the Json Query editors I demonstrated previously, but so much easier to use I much prefer it.

Any modifications? Improvements? Bugs? Please send them through. I've used it on lists up to 50,000 members. Chrome is my browser of choice, but even IE can handly fairly large lists.

Try NimbleSet Now

A site that maintains all the world's 'lookup tables.'

Every lookup table you'll ever need, exposed in open, compatible and readily-consumable ways.

All lists can be retrieved as html or xml, as json, as csv (the one format guaranteed to outlive the cockroach), or any other commonly used** data format.

Lists at 'wiki lists' are more complete, more accessible, and much lighter-weight than the Lists at wikipedia.

The lists are crowd-sourced, contributed by anyone and curated by the detail-oriented masses. The faceless trainspotters and plane spotters, the aspies and cybrarians.

Lists that are currently shown in haphazard ways inside wikipedia could instead source their content from such wiki-lists.

Where a list can be reliably sourced from a canonical source, it is done so programmatically and the attribution provided in a standard way.

Example lists include...

• Every sovereign nation
• Their dialing code, currencies, major languages
• The three letter codes for every airport, landing strip and helipad in the world
• The name and abbreviation of every state in the united states (and every other nation)
• Zip codes and postal codes
• Monarchs of England
• Breeds of dog
• Keywords in C
• Mayors of southhampton
• Notes in F major pentatonic
• Winners of the Belgian grand prix
• Stocks that trade on the NYSE

And a thousands other reference lists. Short lists. Information that changes slowly, slow enough for humans to curate it.

And they're the kind of lists that you (or someone like you) needs to reference in their applications and research all the time.

(See also this thread at Hacker News which prompted me to write down this idea I've thought for so long.)

** Note I said 'commonly used' rather than 'standard'. CSV is commonly used, but it doesn't have a standard. Oh you might point out that there are standards for CSV, and of course there are. Where there's standards, plural, there is no 'standard'.

When you've reached the limits of Haskell on a Horse...

If FubuMVC is not for you...

It's time to meet the MVC web framework to end all MVC Web frameworks.

DOS on DOPE is the modern MVC framework built on the awesome power of Batch scripts.

All of the controllers in DoD are batch files.

All of the views are batch files.

The model is based on batch files. The helper functions are... you guessed it! Beautiful batch files!

The complete Source code and tutorials are available at DoD.CodePlex.com

With DosOnDope, it's this simple:

open a cmd window, and write:

dope site Blog

You just built a website, called 'blog'. Here it is:

Oops. It moves too fast, I can't keep up.

Here's all the folders that have been created for us...

C:\temp\blog>dir /s/a:d/b
C:\temp\blog\Content
C:\temp\blog\Controllers
C:\temp\blog\h
C:\temp\blog\Models
C:\temp\blog\Scripts
C:\temp\blog\Views
C:\temp\blog\Controllers\Home
C:\temp\blog\Views\Shared

And all the files...

C:\temp\blog>dir /s/a:-d/b
C:\temp\blog\Default.aspx
C:\temp\blog\favicon.ico
C:\temp\blog\Global.asax
C:\temp\blog\web.config
C:\temp\blog\Content\Site.css
C:\temp\blog\Content\li.png
C:\temp\blog\Content\Logo.png
C:\temp\blog\Controllers\_actionMissing.bat
C:\temp\blog\Controllers\_controllerMissing.bat
C:\temp\blog\Controllers\Home\Index.bat
C:\temp\blog\h\a.bat
C:\temp\blog\h\h1.bat
C:\temp\blog\h\h2.bat
C:\temp\blog\h\h3.bat
C:\temp\blog\h\p.bat
C:\temp\blog\Models\split.vbs
C:\temp\blog\Views\Shared\_header.bat

Let's create a new controller action, and see how it looks.

Let's improve the look of that page. You type a little...

...hit refresh and see what changed.

It worked.

Notice the use of html helper methods like 'p.bat' and use of the partial, '_header.bat' to give styling and a title to the page.

But websites don't live in a vacuum. We need some data to bring this thing to life.

In DoD we embrace the NoSQL movement and jump straight to the data-store of the future: a CSV file.

C:\temp>CD Blog\Models
C:\temp\Blog\Models>copy con Posts.csv
ID`Title`Article
^Z

Posts.csv is both the model that describes our data, and the store which will contain the data itself.

Let's generate actions for administering our Posts:

C:\temp\blog\Models>dope gen Post
Generating actions for MODEL Post in Controller Home...

Check in a browser...

The generator has created a nice page for us, with a name and textbox for every field, so we can create new posts.

Once we've added a few articles, let's see what the index looks like:

If we look inside the posts.csv file, we see the data:

C:\temp\blog\Models>type Posts.csv
ID`Title`Article
1`Hello Brazil!`Yes, yes, hello indeed!
2`Let's add a second for good measure`Oh yeah

Frequently Asked Questions

Does it scale?

So far I haven't run into any performance problems. Whenever I've opened up a DoD website to several users, my hard drive tends to get wiped long before I discover performance issues.

What about security?

If you can't trust public visitors from the internet, who can you trust? Executing user input as commands inside the operating system is the most direct way to get things done, and that's what matters most. Isn't it?

Well, that's about it. Source code and tutorials available at DoD.CodePlex.

A full-featured web development framework with the smallest possible concept count.

Feedback welcome.

Since then I've researched a bunch of other JSON Query Languages, and created similar editors for each of them.

Here's the full set:

• An editor for SQLike
• An editor for JsLinq
• An editor for JSinq
• An editor for JLinq
• An editor for JimmyLinq

What are they?

Plus, Minus, Interesting

Here's a quick PMI comparison of each of them.

Some other JSON Query languages I could've looked into if my travel time was significantly longer:

• TrimQuery
• JsonQuery
• Jaql

All of which would stop me from getting back to releasing the product I should be releasing, or any of the other technical obsessions I'm trying to develop. So if you're interested in any of those... you're on your own.

You wake up, in unfamiliar surroundings.

You must've experienced some kind of memory loss because you have no idea how you got to where you are. Your forehead is brittle with dried blood.

In front of you is a computer screen, and on the screen it says:

a = Add(2,3);

//a is 5.

b = Add(2);

//what is b?

Intriguing. What is b?

Because you're uncertain of the language in question, you can't be sure.

It could be that Add is an overloaded function, in which case b would be an integer. Its exact value is anyone's guess.

Or it could be that Add has an optional second parameter. Again b would be an integer, though its value could be anything.

It could be that the code refuses to run. The missing parameter is a deal breaker.

But you look at the tattoo on the back of your hand, the 11th letter of the greek alphabet, and you are certain.

What then, is b?

Here's a SQLike editor I wrote. It lets you get your hands dirty with SQLike, the client-side query engine from Thomas Frank.

This started a few weeks back. I read that Thomas had created a nifty DSL in Json for querying (and modifying) arrays of Json. The correct name for it would've been 'JQuery' but that name had already been taken, so he called it SQLike. ;-)

I set about writing a simple editor for it (in the style of the World's Simplest Code Generator) and came up with this fairly usable SQLike editor.

I don't assume you've got any Json to query, and instead expect you to start with a chunk of CSV. I turn that into an array of Json (named 'JsonData') then exec your query over it.

The UI lets you join and query up to two tables at once. I meant to extend it to N tables, but bus-time only lets me do so much, and I keep running into interesting people when I ought to be programming.

In my day job I hand write a lot of sql. And with this editor I quickly found that even though SQLike is lighter than XML it's quite 'ceremonious' compared to writing pure SQL.

Apparently Thomas is hoping to making the syntax for SQLike a lot lighter yet, based on comments here including:

"...stop using pipes - the hell with it: reserve a few words :-D..."

...a sentiment I fully agree with. I look forward to any future versions. If you're listening Thomas: keep going! This is an intriguing project!

Also, I want to add I've already found SQLike useful in helping to achieve my day-to-day data-manipulation activities. I can go straight from CSV to 'queries' without having to put the data into a database first.

Sometimes I can't help but picture the future as an endless sea of cubicles, populated by workers struggling to perform data-manipulation duties. And many will continue to get by without any real tools at all. No sql. No sqlike. Not even a wscg or an excel-formula to help them out.

A future "bus project" might be to implement a similar editor for the overlapping project: JsLinq.

Feedback, criticisms, ideas and suggestions always welcome.

Gorn. --> Try it out. <--

Also, deep within the myriad combinations, I've buried a clue about the new micro-Isv i'm building ;-).

HOT GUIDS - Socializing the Guid

Finally there's a place where people can vote on and discuss their favorite guids.

I've built in pretty much all the best social features. You can vote on a guid, see other votes for a guid, adopt a guid, or email a guid to your friends. And if you get sick of the guid you're looking at: with the click of a button you can move onto a whole new guid.

I'm trying to really take the long tail of human interest by storm. If I get just one customer for each guid, then I'll get... oh I don't know, it must be a *lot* of hits.

Thanks to Stack Overflow for the guid generating snippet. The web template started out the same as 'CreditCardOlogy' but with a little (Powerpoint + MSPaint + PngOut) I took it to a new place entirely.

If you're unsure what any of this is about, please read about guids (& UUIDs) at wikipedia, or try running this Simple Proof That Guid is not Unique.

And make sure email me if you see any particular delicious looking guids. There's some beauties out there.