Mercurial and Tortoisehg for the win! It is an easier to use equivalent to git. You may also want to take a look at Kiln which is built on Mercurial and adds some gimmicks like a wiki and a ticket system. I went for a clean Mercurial install on my own server, but I came pretty close to using their service instead. Fossil is also pretty cool if you want to want to explore alternatives...

I would go with Mercurial and tortoisehg instead. I use git at work and mercurial for personal projects and they are essentially the same, but mercurial stays out of my way more... Git isn't bad, but there are usability issues like having to add new files before a commit or having to pull/merge/push instead of having it just create unnamed branches and letting me merge them when I feel like it. I don't really miss having to deal with local VS public branches when using mercurial. And if we could just figure out some way of having tortoisegit remember our usernames/passwords, using it would be much more enjoyable.

Another advantage is that there is no need for backups at all since every developer has the whole repository. I have all 4 source control systems on my server (svn, mercurial, git and fossil). Surprisingly, Fossil was the easiest to setup, followed by mercurial and svn. Kiln is pretty cool if you want a mercurial server without the headache of hosting and the command line server configuration. They offer an issue tracker and other useful tools too. The only thing I miss is a good artifact manager that will fit the branching scheme and really feel integrated in the process.

Well I am one such developer who recommends a dvcs or single man teams. The reason is not the distributed aspect of it though. It's the easy true branching. I always keep a released and a development branch and I can switch to any version easily using a graphical representation of my versionning scheme. I'm not a fan of git because of the complexity and of how easy it is to break things. I prefer mercurial with tortoisehg. Fossil is interesting but last time I checked (a long time ago), it lacked integration with Windows explorer. Whatever the tools, you need a good branching scheme. A dvcs just makes it easier.

I just used zombie mode yesterday. I basically spend hours running in circles and barely moving forward tough! It's nice to get an energy boost, but zombies are clumsy...

I congratulate your multitasking efficiency : We should hire more men like you.

ChrisElston wrote:

Of course the grandkids had long got bored watching me faff about with it and moved outside or into the other room to do something else.

They were just testing distractions for the next time you must watch them.

I got 4x1080p monitors, two horizontal, two vertical. Verticals are for code (comparing two files with this is awesome) and browsing (the 1080 width is rarely a problem, but it happens). When you get to coding on vertical screens, it will hurt to go back. One of the horizontal ones is for the palettes from whatever editor that supports placing them on another screen. It may also have a music player, IM, etc. The last one is for porn.

The mind often finds solutions to the problem at hand by wandering in totally unrelated fields... How many times have you gone to bed only to think of a solution to some problem you had during the day? I think the same happens on a lower scale. If I try to concentrate on the problem, I can make logical conclusions towards a solution. But if I let my mind poke around, I can be creative and invent new solutions that I wouldn't have thought of otherwise. Or I might just waste a few hours thinking about how generalized anesthesia works (it's quite interesting actually).

I found this last week.

// Remove this code.

Followed by a few lines of unnecessary code that had no effect in the context. And source control says the last modification on that file was in 2007! :doh:

I started a part-time WPF project a few months ago and the learning curve isn't so bad. The first week wasn't very productive, but I've managed to develop a C-like code diff tool (Beyond Compare + syntax coloring + better line matching algorithms) with custom scrollbars, document coloring and stuff very easily. I went through a similar learning curve in ASP.NET and it never felt as natural as WPF. But I hate Javascript so maybe I'm biased...

I think you need to make it goal oriented... Something like this : 1. Create a console application that prints its first command line argument. (the basics) 2. Make it print all of them and print a warning if no argument are passed. (flow control) 3. Make it parse all arguments as int and sum them. (use variables, understanding types int/string) 4. Make it take the first argument as a mathematical expression, split by operators (+ and -) and apply them. (advanced work with data) 5. Declare an interface Expression that define a GetResult and a class for Number, Addition and Subtraction, then build an execution tree to get to the same result. (basic OO) 6. Make it work using Polish notation instead. (refactoring) 7. Implement Division and Multiplication with correct operation priority. (exploring OO) 8. Create a Windows forms UI for a calculator with a text box where the user can type an equation. (forms) 9. Parse this equation for parenthesis and operation priority and reformat it as Polish notation. (advanced data manipulation) 10. Pass this equation to the command line application previously developed and show the output. (how to look for functionalities in the framework and MS documentation) 11. Integrate the calculation code into the forms application as a separate project, add the dependency and make it work. (multiple projects, dependencies and refactoring) 12. Create a database in SQL Express and add a database for logging previous calculations with input and output. (database design) 13. Use a typed dataset or entities or whatever technology you want him familiar with to add one row to the table everytime a user presses "calculate". (simple inserts) 14. Use the database as a cache to bypass the calculator and retrieve the answer from the database if the calculation was requested previously. (simple queries and refactoring) 15. And from then on, he's ready to start fixing small issues or developing small chunks of functionality in your actual applications, either in pair programming or with constructive code review. The rest is experience! Btw, did you know about this? http://tryhaskell.org/[^]

I've been reading/hearing about GPU based calculation a lot recently. Apparently, matrices can be sent to the GPU and processed in batch to greatly optimize the calculation. Maybe that would be worth looking into?

Love?[^] Watch this, then kill her dog and hang it in her bathroom! I do it all the time, and they ALWAYS call back...

How about you drill a hole in the wall and put standard routers outside so they escape the Faraday cages. All you got to do is build two boxes (think mailbox style, one per building) to protect it from the rain and your two outside routers can communicate without problem. Put one under the patio and the other near the electricity counter or something and MLW won't even know it's there. Just make sure there is no condensation (don't insulate it) and I think you would do just fine.

I use references and set their property to "copy local" or something like that. You can reference other projects so the dll is rebuilt and imported automatically to your build directory. For asp.net, I guess you could use the strategy of having an "assemblies" directory and set their build action to "copy to output directory", and "copy if newer". If that is not available in asp.net, you could use a post build script to copy the files from assemblies to bin. And if all else fails, I think it's possible to declare that directory as a dll search path or something in the config, but that's less than ideal because it breaks the "standard" directory structure of an asp.net site...

No need to filter out, you can put everything on the share directory! j/k :p VisualSvn is great for that, it's a shame you can't use it. In C#, you should ignore these (maybe VB.NET has more?) - Solution.suo which contains "solution user options" - Project/bin/ which contains your builds - Project/obj/ which contains temporary files for your builds Also, the typical SVN structure is to have your application files within the "trunk" sub directory. This is useful so you can have another root directory for documentation, installer files, etc. and so you can have branches and tags directories too. On the other hand, I've just switched to "distributed version control", namely Mercurial with TortoiseHg and I strongly recommend everyione that they look into it. I won't try to sell it to you (it's free :doh: ), but if you're changing for a version control system, it probably is the best time for you to evaluate it. Even if you don't go for it, it will help you understand the version control concept in general. And there are many advantages to Mercurial like local commits/branches and a "real" tag feature. Here's an article I read recently (thanks to the CodeProject daily newsletter) about structuring a Git repository (it's the same as Mercurial, but it was designed by some operating system inventor or whatever ;) ) : http://nvie.com/posts/a-successful-git-branching-model/?[^]

You can just configure the IDE to replace tabs with spaces whenever you type them... Then again, it can do most of the indenting on its own too!

I agree, you need to determine your own requirements for this software first. Since I don't habe any large scale experience with requirements gathering, I'd start working on it with basic tools to (beginning: ) adjust the requirements for this software, then jump to a better tool (goto beginning; :-D ). I guess the most difficult part in this is to transfer from one application to the next. Anyway, let us know, I'm very interested in what you find, and more importantly, the decisions you make to get there. Good luck!

I've worked in the field before and I'm doing said degree right now and believe me, if you're talking "getting things done", you'll learn much faster if you just get a job in the desired field. And read some good books like code complete. I'm getting that degree for many reasons including that I would love a job in programming language research (my final project is going to be awesome!), etc. Anyway, if you don't like it there, it's a good enough reason to leave. All that really matters if that you enjoy yourself. If you start compromising on that, you'll be doing compromises all your life. The simplest rules make the most efficient algorithms ;)