Re: [despammed] Re: Linux-to-Linux copying?

["Stephen J. Turnbull" <turnbull@example.com>]

>>>>> "Stuart" == Stuart Luppescu <s-luppescu@example.com> writes:

    Stuart> I still don't see what his personality has to do with the
    Stuart> software.

A lot.  Just as culture does.  Eg, there's a saying about development
groups: "a software development group is doomed to replicate its own
table of organization in the design of the software."  My experience
with Japanese OSS (kterm, kinput2, XIM, Mule, nkf, the infamous Pine
patches, Japanized Ghostscript, Japanized TeX, and I'm listing only
the ones that I have read source code for) indicates that Japanese
modules tend to interact with each other more behind the scenes (eg,
using global variables, and those often defined at point-of-use) than
the related (mostly American) code.  (Not to mention being even less
likely to correspond to or even have a design document.)  Just like
their organizations....

Or (a real example) consider a fastidious "Felix Unger" type who
doesn't hesitate to impose his ideas about correct coding style on
code designed by others, even though doing things like s/int/size_t/
for loop indexes, array sizes, and array offsets implies a semantic
change.  (A semantic change which cost yours truly 7MB of mail,
straight to /dev/shredder in one C-x s.  Hm?  Of course, I had
backups!  Don't you?)  Yes, personality matters.

More generally, some people are good designers, others good coders,
yet others good debuggers.  Dunno how those are related to one's AssQ
but Felix Unger-ness definitely bumps it 20 points or so in my book.
(Not to deprecate clean code---if I weren't personally so pissed off
about my mail I might be willing to consider that the net effect is
positive, even, cleanliness is that important in a big project---but
cleanliness at the _expense_ of correctness is not normal.)

    Stuart> Does the fact that Theo is an asshole make OpenSSH lower
    Stuart> quality?

It depends on how he is an asshole (if he is, I tend to trust Chris on
the like-knows-like principle---shut up Chris, how do you think _I_
know you? ;-)---but have never dealt with him personally).  Not taking
well to criticism is not something I associate with high-quality code.
Even the best developers can occasionally have a blind spot.  Even the
acknowledged best developer in XEmacs, widely known as a prima donna,
can be criticized if you approach the matter correctly.  In fact, once
you know the key, he reacts rather well to it.

Control-freaking[1] is a good way to decrease quality in a large
project (this is certainly true of GNU Emacs, both the mainline
version and XEmacs); Theo has several on his plate simultaneously.

    Stuart> Is it that much of an issue that you would prefer a
    Stuart> closed-source alternative?

In security, that's a double-edged question.  If by "closed-source"
you are (which is the case here) adopting the Cambridge Catholic
position that that which is not 100% free is "closed," then yes, it
is.  Because in this case the "closed source" in question is freely
available for public perusal.  If the code is better, the license
doesn't matter, if it allows the uses you need.

If you really mean "closed source", then the answer would be "OpenSSH
is preferable, even if Theo were Satan himself."  Unless you can see
real bugs in the OpenSSH and Theo won't fix them.  Right, Chris?

Aside to Chris: Hey - we all missed that one: "Security Through
Obscurity." -> OXYMORONS.  (Yeah, I know, not two words.  Tough ... a
foolish consistency ...!)


Footnotes: 
[1]  Knuth is _not_ a control freak, BTW; he defined TeX using the
TRIP test and then he set all the rest free.  This is a reason why TeX
is so robust IMHO.

-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."