[racket] Racket on Centos?

From: Neil Van Dyke (neil at neilvandyke.org)
Date: Thu Feb 5 16:04:37 EST 2015

Steve Graham wrote on 02/05/2015 02:24 PM:
>    Out of curiosity, what kind of application have you developed for 
> your client?  Did you have to justify your use of Racket?  Why did you 
> choose Racket?

That particular system was originally developed by others, I've been 
consulting to help them evolve it, and... NDA.

There are a various anecdotes and perspectives you'll hear from people 
who developed systems using atypical languages/platforms. For example, 
there are public comments by someone at ITA (since acquired by Google), 
who developed a large system in CL, and who said something like they 
thought doing it without some of the CL features was impossible.  
(Racket happens to share those features, and overall I much prefer 
Racket to CL.)  I think the best reason to choose an atypical 
language/platform is because it gives you a strategic advantage.  (For 
Racket, a likely strategic advantage is its metalanguage strength.)

There is also a funny side benefit in that, if you choose an atypical 
language that some of the sharpest and most knowledgeable people 
gravitate towards disproportionately, this seems to result in you being 
able to hire a caliber talent that you probably couldn't otherwise.  For 
example, it's public knowledge that ITA hired a dream team of noteworthy 
Lisp and CS people, for a project that I suspect didn't otherwise sound 
attractive to most people.  I suspect the attraction for many was the 
chance to do big things with Lisp, or to work with other noteworthy 
people.  If this is sounding similar to some of the attraction of, say, 
Google, I think it is -- only, unlike the "Google" household word, it 
just usually doesn't impress your family and friends (unless they are CS 
people).

If you're trying to get some kind of approval for your platform choices, 
such as from established-company executives or startup investors... that 
can be hard.  You really have to bring the goods, so that the unfamiliar 
seems a worthwhile risk.  And, if you're successful thus far, you might 
have to renew that faith periodically, whenever it's time for another 
dose of architecture investment.

In a different situation, if you're doing moonlighting projects, like 
apps and Web sites, without investors&executives&managers... you can 
have reasons that would be harder sells than obvious strategic 
advantage, like "I feel more productive with it", "it's more fun", 
"parentheses are aesthetically pleasing (though pound-colon is not)", 
and "I can search/ask a question about the platform, and get a high 
signal:noise ratio."

Neil V.


Posted on the users mailing list.