Kyle Cordes

Most of the software I work with is very practical. At Oasis Digital we mostly create line-of-business enterprise software, and even when I step away from that, I usually pick up a tool or language that has a good likelihood of mainsteam adoption.

Sometimes, though, I like to really stretch my mind. For that, it’s hard to beat Factor. Factor is fascinating in that it combines a goal of efficiency and practicality, with a syntax and computation model which are quite alien even to a software polyglot. Don’t let the stack-ness deceive you; it’s a big leap even if you’ve used FORTH and grown up with a HP RPN programmable calculator.

So I set about this evening to work with some Factor code, a simple GUI calculator posted a few days ago by John Benediktsson. I bit off an apparently small bit of work: remove the “code smell” of that global variable, and in the process, make it so multiple calcs each have their own model (rather than a global shared state).

Original version from John

My finished version

The two most important pieces of updated code:

The changes consist approximately of:

• Change all the button words to accept a model input
• Change the <row> word to accept a model and use map instead of output>array
• Remove the calc variable
• Change the calc-ui word to shuffle things around and use make rather than output>array

In case it isn’t obvious from my text above or the source code, I am not a Factor programmer, please do not use this as example code. On the other, I learned a bunch of little things about Factor, and perhaps implicitly about concatenative programming in general, in the process of making this work.

Tags: factor

I adopted Git as my primary source control tool a couple of years ago, when I was using Windows as my primary (90%) desktop OS. Since then I’ve switched to 75% Mac OSX, but I still use Git on Windows for a few projects, and I get a lot of questions about Git on Windows.

I use msysgit (and its included GUI) most often myself, but I don’t have a clear answer as to which is the “best” Git GUI for Windows. I can offer this list of choices, though, along with some thoughts about them.

There is also a very long list of Git tools on the main Git wiki; but that page is just a list, without any other information.

msysgit

msysgit is the main project which ships a Windows port of Git. It is based on MSYS, so it fits in the Windows ecosystem a bit better than the cygwin Git port.

msysgit includes the same Tk-based GUI tools as Git on Linux: a commit tool and a repo-browse tool, plus a bit of shell integration to active the GUI by right-clicking in Windows Explorer, plus a new thing call git-cheetah, which appears to be heading toward Tortoise-style integration. These tools are a bit ugly, but have good and useful functionality. I don’t mind the ugly (I get my fix of stylish software over on my Mac…), and I find the features ample for most work.

If you don’t know where to start, or if you want a Linux-like Git experience, start with msysgit and learn to use its tools.

msysgit is free open source software. It is under active development, and keeps up with the upstream Git versions reasonably well. There is even a portable (zero-install) version available.

My biggest gripe with msysgit (and its GUI) is that I had to figure out how to use it effectively myself. I could have really used a video walkthrough of how to be productive with it, back when I was starting out. That was a long time ago for me, but might be Right Now for people reading this post. If anyone can suggest such a tutorial, I’ll link to it here.

TortoiseGit

This is an attempt to port TortoiseSVN to git, yielding TortoiseGit. If you like and use TortoiseSVN, you’ll probably find this worth a try. I haven’t tried it yet myself.

TortoiseGit is free open source software, and is under active development.

Git Extensions

This Git GUI has a shell extension (like the Tortoise family) and also a plugin for Visual Studio. From the screen shots, it appears to be feature-rich and complete.

Git Extensions is free open source software, and is under active development.

SmartGit

Unlike the other tools listed here, SmartGit is a commercial product (from a German company), starting at around $70. It appears to be more polished than the others, as is often the case with commercial products. It also appears to be quite feature-rich.

I don’t know how SmartGit fits in with the Git licensing; Git is licensed GPL (v2), so I assume (hope?) SmartGit has found some way to use it under the hood without linking to it in a way that would cause license trouble.

SmartGit requires a Java runtime, implying that it is written in Java. Five year ago I thought of that as a caveat; but today, Java-based GUIs can be extremely attractive and fast, so I don’t see as a problem at all.

Is IDE Integration Vital?

I know people who swear by their IDE experience, and are aghast at the thought of any daily-use dev tool that is not integrated with their IDE. It is almost as though for this group, multitasking does not exist, and any need to run more than one piece of software at the same time is a defect.

Now I love a good IDE as much as anyone (I’ve urged and coached many developers to move from an editor to an IDE), but I don’t agree with the notion that source control must always be in the IDE. IDE-integrated source control can be very useful, but there are sometimes cases where non-integrated source control wins.

The most common example for me is when using Eclipse. There are two annoyances I see regularly:

• Eclipse assumes that one Eclipse project is one source control project, an assumption that is sometimes helpful and sometimes painful. In the latter case, simply ditch the Eclipse integration, and use a whole workspace (N projects) as a single source-control project, outside of Eclipse.
• Sometimes Eclipse source control integration bogs down performance. Turn it off, and things speed up.

Therefore, when I use Eclipse, I generally manage the files from outside, using msysgit, command line, etc.

Feedback Wanted

I’d love to hear about:

• More Windows Git GUIs to list here
• Anything else I’ve missed

.. via the contact page (link at the top of the page). I try to reply to all email within a few days.

Tags: git, source-control, windows

Over the last year or so, I’ve spent perhaps 50 hours rethinking what kind of business Oasis Digital should be. I’ve studied business models. I’ve made spreadsheets. I’ve looked around that numerous other consulting firms. The results of all that… are slowly emerging. Stay tuned.

In the meantime, though, I noticed something very interesting: the firms that appeal to me most, in terms of web site content, community involvement, portfolios, marketing approach, etc., are disproportionately Ruby or Ruby-Rails shops. I admire the “vibe” of the Ruby community:

• Strong focus on design, to the extent that some Ruby-centric development firms have web sites which could pass for visual-design firms instead.
• Ruby developers seem unusually aware of the extent to which syntax and conciseness matter.
• There is much discussion of craftsmanship, though I’d need to survey a broader swath of production code to determine whether this discussion has a basis in reality.
• Seemingly contrary to the above factors, Rubyists also appear to be unusually pragmatic.
• This pragmatism translates to real-world financial impact: many developers make a good living with Ruby, and many firms are very happy with their Ruby projects.
• Ruby events are numerous, nationwide.

There are numerous Ruby- or Ruby-Rails-centric development firms, and Oasis Digital is not one of them (we are perhaps a 5%-or-so Ruby shop, with Ruby expertise to effectively attack automated sysadmin, integration projects, and so on). We aren’t going to become a Ruby-centric-firm, either; and there are some technical aspects of Ruby that don’t impress me.

Rather, we want to bring some of the cultural qualities seen in the Ruby community, to other languages and tools. We care about design much more than most firms, and it shows in our GUIs. We care about user experience, and we are obsessed with quality, working results.

Tags: commentary, ruby

Recently I analyzed the relative payoff from different types of work I’ve done in my career to date. Some of the work has paid off reasonably well. But one particular bit of it stands out as a counter-example to common wisdom:

Between 1997 and 2000, I spent countless hours on the BDE Alternatives Guide, a section of this web site devoted to listing and analyzing the dozens of third-party database access libraries available for Delphi in that era. Delphi shipped with the BDE a not-great mechanism for database access. BDE was Borland’s answer to Microsoft’s ODBC, but unlike the latter, BDE didn’t get industry-wide support.

Working on the BDE Alternatives Guide had many positive payoffs:

• It created a much-needed resource, greatly appreciated by thousands of developers.
• I learned enormously in the process.
• It put me in touch with dozens of library vendors, and many hundreds of developers.
• It generated many incoming links and much traffic, around a million page views over a 5-year period.
• Banner advertisements brought in a few hundred dollars, before I scrapped them to avoid the appearance of bias.
• It made me reasonably well-known in the Delphi world, which was growing rapidly at that time. (Our team at Oasis Digital still does some Delphi work, by the way.)

You might think, though, that establishing expertise as a Delphi database integration expert, would result in lots of consulting leads, new business, and job offers. Let’s look at the stats:

• Total number of leads generated: 0
• Total consulting work generated: $0
• Total job inquiries and offers: 0

Yes, that’s right. Not a single firm ever contacted me to inquire about consulting, development, etc., as a result of the BAG. I’m still glad I did the work, for all the reasons above. But it is a counter-example to the notion of showing expertise and building a following, as a way to generate business interest. Sometimes the work pays off in new business, and sometimes it doesn’t.

Tags: commentary, delphi

Node is a combination of Google’s V8 JavaScript implementation, and various plumbing and libraries. The result is an unusual and clever server programming platform. Node is in a fairly early development phase, and already has a remarkably active community: ~9000 mailing list messages (as of June 2010) and many dozens of projects and libraries. I’ve spent some time digging through Node code and writing small bits of it, and was pleased with what I found.

Why is Node worthy of attention?

• JavaScript is a Next Big Language, it is everywhere. It is probably the most widely used programming language ever.
• I know a few things about asyncronous server programming, having done a lot of it in 1990s IVR software; it is very well suited to serving a large user population.
• Node is accumulating libraries at an impressive rate, indicating momentum.
• There are significant advantages in developing a whole application stack (server and client code) in a single language. For example, this makes code and business logic sharing works across tiers. Using Node, a JavaScript-HTML tool, a JavaScript-CSS tool, JSON, etc., it is possible to develop a complex web application using only JavaScript.

Node is not all unicorns and roses though.; my most serious misgiving about it is that it does not (yet) have a great strategy to make straightforward use of many-core servers. We’ll have to see how that develops over time.

Node Knockout

The team at Fortnight Labs is putting together Node Knockout, a 48-hour Node programming contest. I am a fan of such contests. I’ve offered to help out by being a judge, and I’ve also signed up Oasis Digital as a sponsor.

As a judge, I can’t be on a team; I’ve like to see a team or two form here in St. Louis, though.

Tags: contest, javascript, node.js, programming

For today, a “day job” topic:

Oasis Digital (my firm) is a custom software development shop. It is not a staffing or recruiting firm; there are many good firms in those businesses, and I have no desire to join them in that market. Oasis Digital doesn’t offer contract-to-hire, it doesn’t charge a percentage of a hire’s pay, and does not recruit in for customer placement.

Nonetheless, we do occasionally help our customers hire.

Really?

I’ve heard customers and our own team members express surprise at this. Isn’t it against our own interest to help a customer with a direct hire, who might end up doing some work instead of Oasis Digital doing it?

In a short-term, trivial sense it is perhaps against our interest. If we wanted every project to go on forever, using 100% only Oasis Digital staff, then we would make sure to never help any customer with any hire at all. But that is completely unrealistic. There are millions of software developers (and sometimes it seems almost as many software development firms). We are in a competitive market. Our customers have a choice, they can do business with us or have someone else write their software instead, regardless of whether we help them with hiring.

Therefore, the real question is whether to be greedy for the short term, or visionary for the long term. We choose the latter. Our policy is that we are happy to help our customers hire direct staff. We believe that this will, in the long term, lead to success for our customers and for Oasis Digital.

We assist with hiring in 3 ways.

Direct Assistance with Hiring and Onboarding

At Oasis Digital we have a somewhat unusual hiring process: in addition to the usual interviews by phone, in person, and otherwise, we ask (and pay) prospective developers to write some code for us based on a short specification. The resulting code, and conversation about it, provides a great opportunity to get to know someone (and to assess the results they will create) very quickly before hiring them. We assess technical skills as well as teamwork / cultural fit. We have a high bar to hiring and a defined process to reach that bar. At the same time, our process respects potential employees, by not asking for sample work to be done for free.

A good hire, though, is not the finish line. It is the starting line! During the first months of a new developer’s work we have an onboarding process in which the new developer sets up a work environment (mostly by referring to project documentation), then implements tiny changes, then small changes, then medium changes, then finally can begin work on large, important tasks. Throughout these initial months, the new developer works with more frequent collaboration and code/change review than will be needed in the long run. We have found that with our hiring and onboarding processes (described above at a very high level), we have a high success rate.

The first and most direct way we can assist our customers, therefore, is to simply execute these processes for them: assist with interviews, sample projects, and lead the onboarding effort.

Same Standards

When working with a mixed team consisting of Oasis Digital staff as well as customer staff, we hold everyone to the same high standards.

I’ve seen teams that work the other way: accepting a lower standard of work from a customer’s internal staff. It ends badly. We would rather lose a customer, than ship bad software. Our reputation matters more than the next dollar.

Pass It On

Lastly, our processes aren’t a deep secret; the key is not the ideas, it is the execution. We are happy to teach our way of working to customers (and everyone else, in blog posts and talks). Even at the price (free to read, normal billing for customer work) it is a hard sell, though: hiring is often deeply embedded in how companies work.

Stay Tuned

I’ve summarized here at a high level; expect future posts and talks with many more details.

Tags: hiring