From: John M. Adams
Subject: Updated Space Telescope language report
Date: 
Message-ID: <oqa8yqylciu.fsf@RAKTABIJA.stsci.edu>
A while back, I posted a report of a language study we did here at the
Space Telescope Science Institute.  It was long (90 pages or so).

I have since rewritten it.  This one is more concise (only 20 pages),
presents a firm recommendation (the former was ostensibly agnostic
according to the disposition of a certain committee), and is more
systematic in presenting the reasoning behind the recommendation.

The report is available here:

http://www.ess.stsci.edu/psdb/spike/spike-wg/psflang/lr3.pdf

The political history of these revisions is that a higher ranking
committee didn't like the lack of a strong recommendation in the
original report.  A second report was written, but that was considered
by senior managers to have an insufficiently objective appearance.
The third report is the final one.

-- 
John M. Adams
From: Andrzej Lewandowski
Subject: Re: Updated Space Telescope language report
Date: 
Message-ID: <ci9bhvkj7kjq31h3urrm70pqacha5ei997@4ax.com>
On 16 Jul 2003 11:25:13 -0400, ·······@stsci.edu (John M. Adams)
wrote:

>A while back, I posted a report of a language study we did here at the
>Space Telescope Science Institute.  It was long (90 pages or so).
>
>I have since rewritten it.  This one is more concise (only 20 pages),
>presents a firm recommendation (the former was ostensibly agnostic
>according to the disposition of a certain committee), and is more
>systematic in presenting the reasoning behind the recommendation.
>
>The report is available here:
>
>http://www.ess.stsci.edu/psdb/spike/spike-wg/psflang/lr3.pdf
>
>The political history of these revisions is that a higher ranking
>committee didn't like the lack of a strong recommendation in the
>original report.  A second report was written, but that was considered
>by senior managers to have an insufficiently objective appearance.
>The third report is the final one.


Great comparison:

1. We selected Java, Python, C++ and Lisp,
2. We rejected Python and C++ up front. On the basis of "anecdotal
evidence",
3. We described Java. But we don't have experience with Java,
4. We described Lisp. We have more experience with Lisp than with
Java, therefore we can write more about Lisp than Java and more
favorably,
5. We wanted to be fair for Java: Java has better GUI than Lisp,
6. We concluded that we prefer Lisp because our project is in Lisp.


A.L.