Chad Perrin: SOB

5 September 2006

another point in favor of “do it yourself”

Filed under: Geek,Metalog — apotheon @ 06:00

As time passes, it gets progressively easier to do things yourself with computers. Even better, it’s getting easier to do things well. This is in large part due to a general advancement of technology, of course. More specifically, it’s mostly a result of two things: open source software development and the Web. Both of these processes (and yes, the Web is essentially a process, not a “thing”) create distribution, inclusion, and accessibility to the technologies they target. In a sense, it’s trickle-up economics — by leveraging the power of the masses, advancing technological accessibility to create a profitable market niche.

As this happens, the excuses against the “do it yourself” approach wear increasingly thin. Meanwhile, the problems with having others do things for you become more and more painful to experience. For instance, fellow weblogger Sterling recently had an outage because of a moment of incompetence from a service provider. Unfortunately, while all the resources for running your own weblog software are at our fingertips, it’s not so easy for us to run our own webservers with any sort of reliable, reasonable-performance capabilities. It’s getting better all the time, though.

The only excuses for using a feature-strong, functionality-weak blogging service like Blogspot or Livejournal these days are A) complete lack of technical skills and B) trying to leverage the internal communities to build up readership and Google attention prior to moving to a self-run solution so that you’ll have some quick PageRank building potential once you get off those dead-end services. There will come a day when using bargain-basement webhost services for hosting your weblog is almost inexcusable too: when that day comes, I’ll be self-hosting my websites already, telling you that you should follow suit. In the meantime, here I am choosing and modifying my own weblog software, telling any of you on Blogspot and Livejournal who have the ability to do what I’ve done that this is where it’s at.

You want the ability to choose your own antispam measures, add your own functionality, and (perhaps most importantly) move your weblog to another webhost if need be, even if you don’t know yet that you want these things. LiveJournal doesn’t support trackbacks and pingbacks. Blogspot’s antispam measures effectively block legitimate traffic. When you use Someone Else’s Service, they will always make it difficult to move your data somewhere else if you decide you need to because, of course, they want to lock you in. They’re not just locking you into one part of the entire stack of technologies, either: they’re locking you into an integrated Leaning Tower of Tech. In that sort of situation, when one little thing fails you, the whole integrated mess needs to be replaced.

Moral of the story: get the heck off Blogspot or Livejournal if you’re at all serious about your weblog. Do it yourself if you want it done right.

The Ecosystem Metaphor

Filed under: Cognition,Geek,Liberty — apotheon @ 04:34

People talk about “ecosystems” all the time. Lately, because of the intellectual circles in which I travel, that’s mostly in reference to “enterprise software” and things that bear a marked resemblance thereto, at least from certain angles.

In these terms, any new market niche exploited by some hot competitors, growing into a multimillion-dollar industry rapidly, is an emerging ecosystem. Something that was once such a thing, but is now old hat in computer years, is an established ecosystem. Ecosystems, ecosystems, everywhere.

The ecosystem metaphor is actually quite excellent for illustrating points of economics and market forces at work. Something I find mind-boggling is that nobody (for some generalized, only mostly-accurate definition of the term “nobody”) seems able to recognize the implications of applying that metaphor. People talk about software ecosystems in terms of how enterprise software frameworks like J2EE are necessary to address the scaling requirements of the evolving ecosystem, whichever ecosystem it is this time.

. . . and there’s another ballpeen hammer to the head that “nobody” noticed: “evolving”. Evolution is an implicit assumption of the ecosystem metaphor. Without accounting for evolution, your ecosystem metaphor is useless. Evolutionary processes are an intrinsic part of an ecosystem. Strict, anti-evolution Creationists need not apply — if the world is as they claim, the word “ecosystem” is meaningless (so let’s leave that side-issue where it lies for now).

Not only does software evolve: the entire software ecosystem evolves. This refers not only to the application at hand, but to the development tools, the development environment, the company that contains that environment, the population of employees who work with the tools and write the software, the competing software offerings from other, competing business endeavors (and all the constituent parts as for the first company), the customer base to which they (hopefully) cater, and the niche market itself. Even the enclosing, total socioeconomic system evolves, and this has a sort of trickle-down effect on all the parts of your software ecosystem to which I have referred. This socioeconomic system is of course defined (mostly) by the myriad of niche markets, both like and unlike your own, that are constantly evolving. It’s all an immensely complex system that, when left to its devices, dances an intricate ballet like a perfectly synchronized dance troupe of more than six billion dancers.

So what’s the problem with huge, unluggable enterprise frameworks?

Well . . . nothing absolute. They’re just used according to decisions predicated upon assumptions that aren’t based in a clear view of what constitutes an ecosystem.

Case study: Viaweb.

Eminent Lisp hacker, venture capitalist, geek startup mentor, and essayist Paul Graham got together with a partner (Robert Morris) and produced a real mover-and-shaker of a killer app that helped to define the future (now current) landscape of Web-based e-commerce. They kicked butt all over their competitors. Paul attributes this all to Viaweb’s use of Lisp, its bold and forward thinking idea-generation geek startup environment, and other factors that fly in the face of generations of ponderous corporate juggernaut economics. They did so well that Yahoo! bought Viaweb to the tune of $49 million dollars.

This became the Yahoo! Merchant Solutions system that exists today, the largest individual e-commerce platform system in the world. Very little of its evolving codebase, if any, is still being written in Common Lisp today.

This is a sign of several things, perhaps most notably that Yahoo! isn’t hiring people like Paul Graham and Robert Morris (who made Viaweb, and made it a success, in the first place) to contribute to that evolution. Yahoo! is hiring what huge corporations’ middle managers always hire: the “safe bet”. They want to freeze the dynamics of the ecosystem at current state, which has Yahoo! at the top. Unfortunately for them, that position of eminence is beginning to erode. What went wrong?

That’s the end of the case study. Let’s examine the lesson from this, now.

The world moved on. New, dynamic up-and-comers are popping up all the time. The Web is still a relatively free market, and allows a heck of a lot of room for startups to take a shot without utterly destroying the lives of the people taking those shots. There’s nothing to lose, as long as you’re not someone like Yahoo! (which has only its eminent position in the market to lose, for now).

In an ecosystem, something like Viaweb is an organism. It comes into being, it evolves, it flourishes, it changes — ultimately, it grows old, is overtaken by younger efforts, and dies. Yahoo! has it (under its new name) on life support and steroids, so it will not give up the ghost and still looks healthy. Given time, it will fall by the wayside and become irrelevant, unless something is done to inject some actual youth back into the thing. That’s not a problem, though, unless you’re a Yahoo! stockholder or employee banking your entire financial existence on the ability of a corporate behemoth to maintain a dominant stasis in what is naturally a very dynamic system. Stasis in dynamic systems has an official, descriptive name, by the way: stagnation.

It’s only the masses of money accumulated by Yahoo!, and brand recognition, that will sustain it against the threat of smaller, more effective competitors. It is only governmental interference in the market in the form of corporate law that allows Yahoo! to sustain that financial solvency and manipulate market forces in the short term to eke out a few more years of market dominance for major business components like Yahoo! Merchant Solutions. Without that, it would be brought down by younger, better competitors who would simultaneously fill its niche and advance the state of the art, even inventing new niches that are more efficient at generating wealth in a socioeconomic system.

If you want advancement, innovation, and real wealth generation that positively effects the entire socioeconomic system worldwide, let the ecosystem do its work. Let evolution happen. Let the strong flourish and the weak perish, no matter how much brand recognition the weak organisms may carry.

Paul Graham, by way of a newish sort of venture capital firm called Y Combinator, is contributing materially (and actively) to enabling the natural processes of software ecosystems, providing a dyke of sorts to prevent the sea of corporate power from drowning evolving new innovations before they have a chance to grow into real threats — and, simultaneously, into benefits for everyone in the market (note that I do not consider a collection of incorporation documents to qualify as part of “everyone”). Huge corporations aligned toward market dominance practices intended to ensure continued survival despite their corpulent bloat, on the other hand, are the old has-beens that would best serve us by getting replaced.

Death is part of an ecosystem. It’s a cycle of life for a reason: it just makes good economic sense.

RMS had RMS: the irony of Esc+Meta+Alt+Ctrl+Shift

Filed under: Geek — apotheon @ 02:25

Does anyone else find it terribly ironic that Richard Matthew Stallman (aka RMS), inventor of EMACS (known for its extremely complex keyboard shortcut sequences), developed a crippling case of repetitive motion syndrome (aka RMS) in his hands that was so bad he was unable to write code for several years? Some might call that poetic justice. I just call it a warning to the wise, and ample justification for my preference for Vim over EMACS.

On the other hand, I really do wish Vim was written in Lisp, and used Lisp for its configuration files, the way EMACS does. Perhaps that’s a software fork project for the (distant?) future.

All original content Copyright Chad Perrin: Distributed under the terms of the Open Works License