The Viamentis Blog

Well, long long ago, Reddit used to run on Lisp. Unfortunately, due to some problems with working setup, they had to move to Python, after which a huge brou-ha-ha followed. Just do a google search, and you will find it. Somebody recently asked on c.l.l as to why the switch happend at all. Steve Huffman (one of the Reddit guys) says so:

“Python drives me nuts, and I miss Lisp, but I’m not the only
person working on reddit, so we use what we can collectively use best.”

Sigh. Yeah. I understand what he is feeling. Lisp kinda grows on you. You can just drop any other programming language you picked up and move on to others, but there’s something about Lisp that you will always miss in others. “Quality without a name”?

Yes, I do think so. There are somethings in life which have that. For me, a BMW car has that (a Merc is somehow too flashy for my taste), so does Emacs (am I mixing metaphors too randomly?) and so do some of the retail stores I admire. They somehow get the right mix of elegance, functionality and price. And it becomes irresistibly appealing.

I was struck by the timeliness of this essay by Martin Fowler. It’s something I face everyday as I’m stumbling through my way to do a simple webapp in Lisp. Well, Ruby’s ravine is deceptively simple, so I might have forgotten the trials of learning a new language. It’s always the same. But for me, the ravine seems especially deep and long compared with say, Rails because it’s optimized for web development, and let’s me prototype something non-trivial in a matter of few hours.

Not so with CL, as I have to cobble together different alternatives for web servers, try them out, different html-generating frameworks, again check them out, and finally whether to go for a JavaScript/CSS-generating framework or not, etc. It temptation is is very much there to just do it in Rails and be done with it. But CL keeps beckoning me with that “Quality Without a Name”. For now, I’m plodding on. Let’s see how far the plateau is, or if it exists at all (at least with respect to webapps).

It’s too early to say, but if they keep building such kick-ass b0xen, why not? Sun has always been a hardware company, and it’s nice to see it launching some really cool stuff. Check this puppy out. Whoa. It’s gotta be a sysadmin’s wet dream. Seems like their machines are getting some influential people excited – among them Joyent, the people behind TextDrive. Go Sun!

Some of you must’ve alredy read the latest Paul Graham missive where he lists out 18 mistakes that are deadly for startups. I don’t have enough experience to opine on all of them, but I do agree with most of them. They seem sensible enough. Except the very first one: Single Founder. Gee. Viamentis sure has one founder (and one employee, so far – he he). So that puts me in a precarious section, I guess.

I do have to agree with some points he makes. The stress is tremendous. And the temptation to quit or take easy routes is also quite incessant. But I don’t think it’s impossible. I don’t know about “startups” the way that PG defines them (he says only product companies can be called startups, which I think is silly), but if you look at other fields other than IT, there are lots of companies founded by a single person. Agreed, it’s much tougher to do it alone, but hardly not doable.

I can even list some of the advantages: no infighting between the founders. The company would have a single vision to drive it, so less looking over the shoulder, second guessing the others, and turf wars. Of course, PG sounds much more convincing with all his experience. Well, all I can say is: we will see.