[linux-elitists] Re: Yet another mozilla atrocity
Mon Oct 13 22:30:39 PDT 2003
Quoting Martin Pool (email@example.com):
> Now you sound evasive.
By impugning my motives, you have just managed to debase this
conversation even further -- congratulations! -- and I will probably be
giving up on it once again.
> You said that you wanted to
> refute an allegation that GConf is a fine way to store application
> I did read that as saying gconf was bad. What were you trying to say?
You know, actually paying _attention_ to context is a really good way to
glean meanings. As I said in the original post, I was responding to a
post in another forum that included this snippet:
GNOME's configuration system is a bunch of text files that you can
edit in vi. Yes, text files. They're XML, and human-readable. GOOD.
In the other forum's discussion, the poster to whom I was responding
(Peter Whysall) was dismissing claims that GConf configuration storage
is alien to Unix traditionalists -- claiming mere use of XML flat files
should make such users feel entirely at home. I was examining whether
his claim held water, finding that it did not, and calling bullshit on
I presented a brief excursion by me, a longtime Unix user, trying
honestly (but unsuccessfully) to test his claim by trying to find and
edit Galeon configuration in the way he claimed would come easily and
naturally. I ended up failing to even manage this with the system-wide
app configuration, let alone the related per-user configuration. (Can
the latter even _exist_ for Galeon? If so, where would the
configuration be stored, and where would I have found that out by
I found no clue in manpages. I found no clue in obvious places to look
in the /etc tree. I found nothing at all in the /usr/share/doc tree
that should have been there but wasn't. (By the way, I'm sorry, but
those omissions taken collectively are simply ghastly, and that is true
without regard to whether I or anyone else send patches. That is a
When I fell back on using locate, I (reportedly) did _not_ find the
configuration data, but rather a red-herring set of XML data intended
for some other purpose.
In some of the follow-up conversation on this mailing list, you told
Karsten that he should have used gconftool. gconftool on my system is a
symlink to /usr/bin/gconftool-1, and there is also a
/usr/bin/gconftool-2. Running either of these seems to do nothing at
all but return a prompt. Neither has (at present) a manpage. Neither
(at present) has a GNU info entry.
So, your suggestion to Karsten that he (and, by extension, I) should
have found and tried gconftool by typing "gconf" and attempting filename
completion suffers the small problem that doing so would have been
completely useless and uninformative.
In any event, the point was not to pass judgement on the storage
mechanism's adequacy from a global perspective -- obviously, I would not
be able to do that from the examination I described -- but rather
whether it passed the usability test that Peter Whysall claimed it did.
In some considerably annoyance, I was venting frustration at its utter
suckitude when examined from that perspective.
The bullshit quotient of the original claim I was refuting has, however,
been handily eclipsed by that of subsequent discussion here. I've been
Cheers, I've been suffering death by PowerPoint, recently.
Rick Moen -- Huw Davies
More information about the linux-elitists