🚀 go-pugleaf

In article <35258@athertn.Atherton.COM> dlw@Atherton.COM (David Williams) writes:

I'm going to respond to the following two articles at the same time.

In case it isn't clear, while I take issue with a few of Garrett Wollmans'
comments, I find David Williams' responses to be some of the most ignorant
and childish I have ever seen on any newsgroup.

   In article <1991May15.175231.1416@uvm.edu>, wollman@emily.uvm.edu
   (Garrett Wollman) writes:
   >In article <35210@athertn.Atherton.COM> dlw@Atherton.COM (David
   Williams) writes:
   >>As for all the anger from the FSF over Apple and look n feel where are the
   >>hoardes of graphical GNU tools of any form? Isn't emacs/epoch the ONLY tool
   >>which "runs" in a X Window? Why aren't all the GNU tools packaged with a
   >>graphical front end? If it isn't important to do it for FSF's own tools
   >>then why the uproar over tools on a platform that doesn't even have a need
   >>for Unix based tools? Unless A/UX is where one wants to put these.

   >Because a "graphical front end" to egrep makes little if no sense
   >whatsoever.  Because the FSF have better things to do, like finishing
   >their free OS and C library.  Because many of the people who use GNU
   >tools have no need or indeed desire for such a "graphical front end"

I've got some definite ideas on making grep (not sure I remember exactly
what differentiates egrep from it) easier to use by putting a graphical
front end on it.  No, experienced grep users aren't likely to need or
want it, at least not at first -- but like with several other tasks I can
think of, tasks where in the past one had to remember "escape" sequences,
I suspect the creation of a graphical front end that eliminated that need
and further provided an accurate and attractive illustration of just WHAT
the pattern will match and/or replace might lure even the hardiest hackers
into the graphical approach.  (Worked with me with, for example, search/replace
commands in simple text editors; not having to decide at the beginning of
the command which character, slash, dot, whatever, wouldn't be in either the
search or replacement string is thankfully a thing of the past -- EMACS
sort of provides this because it is a bit more "graphical" than old
command-style editors.  Even TECO was better.  But being able to easily
mouse-click back and forth between search and replacement string, especially
when the grep feature is available, is the best yet.)

As far as the rest of Garrett's statements in the above paragraph (now off
your screen, perhaps, due to my ranting), I agree somewhat that the FSF
volunteers have "better" things to do, but I'd word it as "higher priority"
things to do to make it clear.  I don't think it is "better" that I am writing
GNU Fortran than doing a front end to grep; but it is currently of higher
priority.  Lots of people want it, and it is considered a more important
aspect of a basic UNIX system than the nifty grep interface (or any other
graphical interface).

I disagree that lots of people don't want the graphical interface, though;
some might not, but I'd say that more importantly they don't want nifty
graphical interfaces for some products at the expense of other required
products, or maintenance on those products, or fast products, or whatever.

Further, adding graphical interfaces to things, especially good ones, is
probably much harder overall than doing traditional character-driven
stuff, but this is ameliorated somewhat by having some control over and
understanding of the underyling software.  Once Project GNU has pretty much
reached the point where the underlying software (the "basic stuff") is done,
adding graphical front ends to stable stuff will be much easier.  Even then,
that sort of work requires (I think) a wider breadth of expertise on the
part of the programmer, something that the GUI-building companies typically
avoid by putting non-programmers on the project.  A few GUI-type things I've
seen built "solo" by programmers looked pretty ugly, even if they were
functional; or were difficult to use, or whatever.  I doubt I'd do any
better.  Look at how much discussion there is over ease of use in EMACS --
this is a powerful system with a programmer-defined interface, many of us
are satisfied with it, but it still isn't considered an "end user" product
by most people.  It's a "step" of sorts from command-driven to GUI-driven
interfaces, being WYSIWYG at the content level.

When Project GNU begins to attract talented designers who may or may not
also be programmers, and is able to have people work together as a team
(already happening to some extent anyway), that'll change things.  In any
case, I think a GUI is simply a much more difficult thing to build,
when comparing breadth of knowledge required to, say, lines of code needed
to make it work, than most anything else on the platter.  (I mean solid,
usable, attractive GUIs -- I'm not talking "cute" but ugly icons, poorly
thought-out designs, and so on.)

   Ah real men don't use GUI's!

This bizarre response by David probably is an attempt to allude to the fact
that, given the choice between a fully featured fast character-based system
like traditional UNIX shells and EMACS, and a somewhat crippled but very
friendly GUI-based system like the Macintosh, many programmers will choose
the former.  I'm one of them, even though I have often appreciated the ease
of use of the Mac.  I'd like to help design the interactions (NOT the
graphics :-) and capabilities for a GUI that gives one direct, visual,
and reasonably friendly access to ALL the power we're used to.  But I'm
probably not competent to do so, certainly less competent than this vs.
doing compilers or OS work.

   Boy did Saber C get it wrong with their product.
   "  "    NeXT    " "          with Workspace Manager and Interface Builder.
   "  "    SCO     " "          with Open Desktop.
   "  "    Hp      " "          with VUE and SoftBench.
   "  "    DEC     " "          with DEC Fuse.

This is garbage.

   >and thus far no one has ever shown the inclination to write one.

   Gee its nice to know such things are unimportant...I guess I'll ring up
   Sun and Unix International, Hp and OSF and tell them to blow off Open
   Look and Motif as they are unimportant and *real* users don't need them.
   Hey I could call up MicroSoft and tell them to recall Windows 3.0
   (3million copies sold), but of course thats merely a layer on DOS and
   therefore ignorable. Hmmm, I could call Apple and tell them that a
   Multifinder interface to A/UX is redundant as well...whoops I mentioned
   the "A" word and a "toy" machine.

Again, more garbage.  Garrett never said or implied "such things" were
unimportant; simply that volunteers have yet to come forward saying "I
really want to write a GUI for something".  Well, I'll step forward and
say it -- but I want to finish GNU Fortran first, and make sure other
more important things (like the kernel) are taken care of, before I do
much work on it.  And I'm probably not a good candidate for it, either.

Another problem: people who really know how to design and build GUIs are
in fairly short supply, and there are tons of highly paying positions
waiting for them, apparently.  The same may be said of compiler/OS designers
like myself, based on the frequency of calls I get from headhunters and
such, but I suspect that if I had the same level of experience building GUIs
that I have had with OS, compiler, and utility design, I might have been
"persuaded" to not volunteer for the FSF.  (Ok, I still would have, but
others might not for that reason.)

   I suppose then one should just toss every other X client other than xterm.

More garbage, with no relation (other than incredibly banal sarcasm) to
anything Garrett said.

Sheesh, when I spent a few months working at the FSF programming offices
doing GNU Fortran integration work, all the other hackers were DEFINITELY
into doing stuff with X.  I don't know much about what they actually did --
aside from setting up my X environment for me and helping me when I screwed
things up (or when somebody else's windows would suddenly appear on my
screen) -- but it sure "looked and felt" like they enjoyed working with
the baby-GUI level of stuff they had under X on their systems.  I don't
know to what extent they're enthusiastic about general-purpose GUIs for
all the FSF utilities, though.

   >Remember:  the FSF is a corporate member of the LPF.  The LPF opposes
   >Look&Feel on grounds of general principle, and the FSF tries to

   Ah, so they're opposed to GUI's just on general principle eh? ;)

I don't remember the context, so I can't be sure why Garrett included this
reminder, but despite the smiley, I can't understand where David made
the connection from being opposed to Look-and-Feel copyright laws to being
opposed to GUIs in general.  I don't think anyone at the FSF is opposed
to GUIs on general principle.  Nothing in any of the postings supporting
the FSF or LPF has even begun to suggest they are against GUIs.

In any case, based on what little I DO know (vs. what little David knows),
I suspect there are a few GUI vendors who are corporate members of the LPF,
disproving any notion that the LPF is opposed to GUIs.

And from what little I know about FSF plans, there's work being done on
making EMACS fit more naturally into an X environment than it already does
(which is to say, not much, except I gather at least it is beyond just
setting up a virtual VT220 or something and "typing" escape codes at it to
move the cursor, based on how it feels when I've used it under X).

   >support this in the best way they can.
   >
   By not providing a GUI for FSF products, I get it.  ;)

This has been said before, but apparently some just DON'T "get it": FSF is
not in the business of NOT providing stuff people can use, unless (for
example) it is something for a system that is the target of a boycott (like
a Mac port of GCC, for example).

And the FSF is not in the business of turning away volunteers for any of these
tasks.  Like doing GUIs for things.

What the FSF is primarily doing right now is Project GNU.  And that involves
creating what is generally recognized as a fully functional UNIX-based system,
including compilers, debuggers, tools, utilities, and so on.  There apparently
are people who think a free "vi" should be part of this instead of EMACS (and
maybe there is -- I know nothing about vi); how much more disagreement would
there be over what is the "best"/"standard" GUI to offer?  That alone makes
the problem a bit difficult.  Even if a volunteer (or a platoon of them)
steps forward and says "I'll do a GUI", the question is, WHAT GUI, what'll
it look like, and how well will it fit with any other GUIs done by other
volunteers for other products?  With most everything else going on under
Project GNU, there is a fairly clear, accepted interface for it, and/or the
programmers feel (rightly or wrongly) they know exactly how to "enhance" or
change it for the GNU version based on their experience.

Next, it is already a little frustrating to not have all FSF software be
as widely portable as UNIX is in general -- e.g. not all systems that can
run UNIX can manage to run gcc or GNU EMACS -- usually because of the
memory requirements.  GUI interfaces will be far less portable, especially
since they'll be based on X (on the other hand, at least for now, X seems
to offer more portability than any other windowing system the FSF could
use -- but more portable designs are possible), so they'll benefit fewer
people.  (At least, this is true at first: from a market point of view,
I'd like to think the availability of a solid GUI-based GNU system would
get a lot more people on the FSF bandwagon and buying UNIX boxes vs. Mac
boxes who otherwise wouldn't choose UNIX systems.  There is a bit of
chicken-and-egg paradox in this; at least with the non-GUI components, we
KNOW people use them and will be generally happy to use GNU equivalents;
with GUIs for UNIX, we can only guess and cross our fingers.)

(Note: when I say "portable" above, I include "practically usable" with
it.  There are, I think, lots of machines that can run typical UNIX stuff
including GNU C and GNU EMACS pretty well, but are effectively useless once
you put X on them.)

Finally, it does make sense for the FSF to have as strong a backing in
the industry as possible before it does something that is likely to get it
sued in a major way -- if they do a GUI that gets Apple or AT&T suing them
before enough people care to take steps to help the FSF out, that'll be the
end of the FSF.  People are likely to love and appreciate the FSF in a year
or two when the kernel is out and stable (I'm guessing at the time frame :-)
and, I hope, after GNU Fortran is out.  (For example, if various oil
companies get high on the FSF because they save lots of money and time by
using GNU Fortran, I'd like to think at least one of them would have a
serious talk on our behalf with any vendor that tries to sue us.)

   >-GAWollman

   David Williams
   "Gosh, and all these Hardware and Software vendors are just not getting it,
    I mean all this GUI work is just getting in the way of writing a free OS
    and compiler/library. Maybe if we sent a terminator back in time to Xerox
    and Apple none of this kind of stuff would be on the marketplace and everyone
    could be using vt100 based workstations."

I normally don't include signatures, but I wanted it to be clear who was
writing what.  That final quote should sum up why I find David's comments
largely useless -- obviously he's decided what the rest of us are thinking
and is ignoring what we're actually saying so he can hurl sarcasm our way.
And certainly he isn't volunteering to help out.

Really, I've never found any posting so childish that attempted to be so
"hip".  Sheesh.  David: Grow Up!  Or at least Shut Up!
--

James Craig Burley, Software Craftsperson    burley@gnu.ai.mit.edu