From: J.C. Roberts (unknownabac.com)
Date: Tue Aug 24 2004 - 12:44:56 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Mon, 23 Aug 2004 16:14:38 -0400, you wrote:

>On Sun, Aug 22, 2004 at 09:40:09PM -0700, J.C. Roberts wrote:
>> If you and your users find the speed of OO or mozilla to be intolerable,
>> there's only two things you can do; (1) use something else and/or (2)
>> work to fix the speed issues. Complaining about it on the OBSD misc
>> list doesn't do any good.
>
>Its sometimes recommended that you read threads that you are thinking of
>replying to. Someone asked why anyone would want to run MS Office under
>wine when OO exists. I told them the reason. So that would be me using
>something else. If you read the thread before telling me not to
>complain, you would realize I wasn't complaining.

It seems you're right on that one. You tend to be rather vehement in
your assertions, so they are easy to misread. On a second read with what
you said above in mind, your statements don't seem as much like
complaints but this is obviously an issue that has really frustrated you
in the past.

>OO and mozilla are both
>apps that demonstrate that being open source doesn't make things good.
>Have fun auditing an app so huge and so messy that bug reports are
>ignored because nobody is able to understand the code that handles that.
>

I think I see your point... -It's mainly a perspective thing. From the
end user perspective, the only thing that matters to them is "speed" on
the usability side. Things like "speed" in the sense of collaboration
with others across platforms are normally over looked by users until
they are faced with those problems.

From the sysadmin perspective, things are very different when your job
is defined as making some solution work across a wide set of various and
incompatible platforms. In this case, the end user's speed complaints
must be ignored if the solution must work across multiple platforms with
minimal administration.

From a security perspective, auditing closed source binaries will always
be far more difficult, time consuming and error prone than auditing
source code. Spaghetti code and poor designs are painful to look at,
worse to work with and even worse to audit but they're still a hell of a
lot easier than auditing a binary. If you want me to bore you with how
some of the MS speed trickery works, let me know off-list but as you
said, to end users, it doesn't matter _how_ the application gets it's
speed and responsiveness.

We agree on one thing; whether the source for a program is open or
closed is most certainly irrelevant to if the program is correct/good or
even fast for that matter. The primary difference between the two is
whether or not you can fix problems or make extensions with a reasonable
amount of effort. Outside of the legal issues, it is still very possible
to fix problems and extend functionality of compiled binaries but it is
far more difficult than working from the programs' source code.

Heck, I personally own over a half dozen Opera licenses for various
platforms simply because it is well written software.

>I simply said real people, in the real world, REFUSE to use
>OO and mozilla because they are so slow. When people voluntarily
>choose to use internet explorer instead of mozilla, its time to admit
>there's something wrong with mozilla.
>
It's a valid point but on the flip side, at least there's a possibility
of fixing things in mozilla and OO. $DIETY only knows what MS would do
to anyone that released fixes/improvements for IE or MSO.

>When you get openoffice running correctly and securely on PARISC-64, let
>us know.
>

The collection of parts for a parisc-64 box sitting on my bench are
(slowly) getting built for testing the OBSD port, so getting OO or
mozilla just _running_ might actually happen but to be fair, correctly
and securely (as well as quickly) are probably out of the question
without significant redesign.

As TedU put it so well, a browser that takes 9 seconds to load can not
be considered correct. There's no debate if some open source programs
have serious problems, instead, it's a known fact that serious problems
exist. Though it's not going to address the immediate end user speed
issues today (or any time soon), at least there is still a possibility
of fixing the problems in the open source programs where the same is not
true in the closed source programs.

Sure, the end user perspective is important but it's not the only
perspective one should be concerned with when figuring out the best way
to deal with these darn machines. I hope I've covered the other ways of
looking at it and why they are important, so I can go back to quietly
testing and improving software in hopes of preventing threads like
this...

JCR

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]