[plt-scheme] Perplexed Programmers

From: Erich Rast (erich at snafu.de)
Date: Wed Aug 29 09:40:29 EDT 2007

>
>> I'll tell you what I think the "software crisis" is.  It's that the
>> difficult stuff is possible but virtually none of the easy stuff is
>> easy.
>
> That's #1. And it is professionals like us who have failed to
> bring across to the general public and to managers that
> programming is not an 'idiot' job (though 'idiots' can write
> things like programs for their own amusement and that nobody
> else uses).
>

Probably counting as one of those 'idiots' (being mainly a hobbyist  
programmer), I find this comment somewhat arrogant and also quite  
misleading.

 From my extensive experiences with it, I'd say that 'academic'  
software (mostly written by computer scientists) is among the worst  
when it comes to reliability and configurability, and I have grown a  
decent amount of frustration about the software and abandonware  
lurking around on university servers. In my experience, 'academic'  
software rarely runs out of the box. Installation and setup tend to  
be extremely tedious. And when you finally get the program to compile  
and run, you still can't use it, because documentation lacks or is  
inaccurate. Contacting developers is quite often not possible or they  
just ignore your decently formulated questions. I could continue this  
list for a long time...

So before pointing at others, I think perhaps CS teachers should  
start to look at their own projects. (I'm not  talking about Matthias  
Felleisen, of course, this is meant to be a general comment---see the  
note about PLT below.) Does my program on the web page still compile?  
Do the server and links work at all? Is the documentation up to date?  
Are the tested configurations specified? Does the web page even say  
what the program does? (No joke, this is sometimes hard to see.) Are  
there test suites? Are all the dependencies noted? Does it run out-of- 
the box? What about bugs? Is there a bug tracking system, release  
notes, etc.?

The good news is that the PLT Scheme suite is a big exception to this  
rule and I haven't got a single complaint about it. Documentation is  
excellent, it works out of the box, it's virtually impossible to  
crash it, etc. But that's an exception---at least in my experience.

Anyway, please don't take this critique personally. I believe there  
is a problem with funding in academic environments. More money should  
be spent on maintaining existing software in academia, and there  
should also be more quality assurance, perhaps even reviews by other  
people from the staff. Academic abandonware IMHO is a big waste of  
money and resources, and just GPL'ing it doesn't make things better.

Best regards,

Erich


Posted on the users mailing list.