Mailing List Archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [tlug] empty rows (PostgreSQL)



>>>>> "Edward" == Edward Wright <edw@example.com> writes:

    >> Hmm ... I see it as relevant, but mostly because this list
    >> isn't really about Linux (which is a kernel) or even about the
    >> GNU System with Linux as the kernel, but about real
    >> sysadmin'ing for the masses.

    Edward> Ok. But how wide do you cast your definition? Even though
    Edward> there is overlap, I don't usually think of web programming
    Edward> as sysadmin'ing.

To borrow Chris's metaphor, if I catch a 3-year-old with a running
chainsaw, then I'm his Dad until I'm between him and the chainsaw
... then we go looking for a _responsible_ adult, hm.  (More likely he
cuts my legs off ... but that's the way it goes.)  The sysadmin's job
is system integrity.  Anything that threatens that, whether it's some
bozo suit with an overfull cup of tea or a buffer overrun in an
extremely popular bozo web programming language or some bozo sysadmin
in .tw with 27 owned boxen (all of them OpenBSD, hi, ayakat & jbyrne!) 
all lighting up syslog like Christmas on every port, that's his job.

    Edward> And while security is certainly relevant here, it's not a
    Edward> security list. We got bugtraq and the like for that. (And
    Edward> the volume there is nearly overwhelming...)

And pretty much noise until you get pretty sophisticated, and even
then sorting out what matters to you requires concentrated attention.

    Edward> <short rant> But newbiews love PHP 'cause it's easy....
    Edward> One of my several beefs with PHP is that is seems to me to
    Edward> have been written from the point of view of providing
    Edward> people who have no idea what they're doing with a bunch of
    Edward> dangerous capabilities. The other languages may be more
    Edward> difficult to learn, but if you make the effort, hopefully
    Edward> you will have acquired at least some "clue" in the
    Edward> process.  </short rant>

Hey, _I_ wrote that ... <checks> oops, sorry again, I didn't.  But I
could have!

Of course to be useful to a new webmin, you need to explain what
"dangerous capability" means.  But it's that kind of rant exactly that
you _won't_ find on bugtraq because it's too "joshiki" to waste baud
on (even on a T3 or better).  And don't worry about not explaining the
first time, because somebody will ask sooner or later.

    Edward> But why don't he talk about that here?

Can't speak for Chris.  His post is a little elliptical, you hadda be
there ... and I'm not into explaining that today (and I missed a bunch
of it, being from way outta town).

    Edward> And why don't you talk about development issues here?

Have done, and Chris is right, that was back in the mid-Morgan and
pre-Morgan era (although to me those are just dates).  I think one of
the things that happened to me was that I picked up a sort of paranoid
fixation on Richard Stallman as the root of all evil (true, but beside
the point, which is that he can't sing) and this really bizarre
hallucination of Jamie Zawinski as the epitome of humor and taste
(true, but he can't sing either, that's why he hires all the great
bands and makes fun of rms), but I realized just in time how boring I
was getting (just in time == before the median tlugger learned how to
use a killfile), and shut up.  There's just so many repetitions of the
twisty prima donna story, all alike that a sentient being can take.

Another thing that happens is that the technical stuff is so much more
specialized, and the basic RPMs all work so well, that you just don't
have to deal with Perl if you don't wanna.  You can be a Pythonista,
and do Mailman ("Rocks! Majordomo Sucks!") or TMDA ("Rocks!
Spamassassin Sucks!") or Zope ("Rocks!" er ... I think the metaphor
just broke down, or something).  But a lot of the stories do involve a
certain level of technical sophistication, and with people getting
ever-farther from the metal and more specialized, you find yourself
talking to yourself quite a bit even with people who _want_ to listen.

Related to that, and one of the things which I think is really cool
but my own project does in a very old-fashioned way is something that
showed up in the original thread.  And that is the way that modern
systems are made up of cooperating agents.  So you've got the RDBMS
and you've got the Apache HTTPd and you've got PHP, and which one is
driving the bus?  Well, that kinda depends on what we wanna talk about
today.  You've got people specializing in the _interface_ between PHP
and the RDBMS, and guess what? not good enough, you need to know
browser too because you need to validate input on the way into the
RDBMS (anything else is asking for serious Mission Impossible-style
trouble) but you also want it in the browser to give the user a
responsive, anticipatory UI---I can spray the buzzwords, as you see,
but one just can't know details of _all_ the languages in the world
anymore.  I don't know these---so I shut up.  But Emacs, well, Emacs
drives the bus or things don't work very well.  And that's not a
modern paradigm.  I'm not ashamed of it---it does its job---but
typically it doesn't provide good parables for these more modern
systems, so I can't even blow smoke when there's no substance.  :-P

Finally, this particular group has moved away from developer-driven
development toward client-driven development, big time.  The social
dynamics are really different.  I got this email today from some guy
"with all due respect aren't all the Emacs and XEmacs developers a
bunch of self-centered wankers?  Doesn't anybody realize that this
lack of unity is hurting your customers == users?"  Well, Guy, ya got
that first part right, but at root (sorry) we're all a bunch of them,
ne?  But ... a _purchase_ is a balanced exchange.  Publishing free
software is a unilateral contribution.  OK, Guy, _read my lips: no
customers, by definition._

It's _all_ about inter-developer dynamics at XEmacs or GNU Emacs; it
can't be all that far from that in the major kernel dev circles.  But
people here are talking about their _jobs_ like all the time.  So some
of the stuff I used to talk about, you know the "herding cats" aspect,
the ESR hacker anthropology stuff, would anybody here understand that?
Dunno, but I don't see people talking about CatB and I don't notice
anybody else missing Hiroo Yamagata, so....

Ciao,

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.


Home | Main Index | Thread Index

Home Page Mailing List Linux and Japan TLUG Members Links