[geeks] And The Linux Weenies Wonder Why They Aren't Mainstream...`

[Frank Van Damme]

On 3/1/06, Jonathan C. Patschke <jp at celestrion.net> wrote:
> On Tue, 28 Feb 2006, Doug McLaren wrote:
>
> Because this mindset is pervasive throughout so many of the open-source
> projects that grew up on Linux.  All the world's an i386 running Linux,
> every compiler is GCC, and if it works on the developer's box and not on
> yours, -you- should fix it.  And if it doesn't so what you want, you
> have the source; why not change it?

It can be hard to test your software under conditions you don't have
access to (like having a dozen of different architectures just to
compile your application on).

> The author of that article just wears that mindset so well.  Silly users
> don't know what they want; they're just sheep.  We should be so kind as
> to let those ignorant morons run our software.
>
> > People just can't look at OpenOffice on it's own merits -- they have
> > to start with how well it deals with Microsoft Office documents.
>
> It doesn't have that many merits of its own.  It's a halfway clone of
> crap software.

And Ms-office is probably a clone of other crap software.

> If I want Office, it's because I need -Office-.  If I
> want to write a document, I have vim and LaTeX.  If I want a database, I
> have PostgreSQL.  If I want a spreadsheet.....well, no one's written a
> halfway decent freeware spreadsheet program, apparently.

You're mixing up freeware and free software?
I am not into spreadsheets, but as far as I know, OOo Calc and
Gnumeric are pretty decent.

> Except that it's slower and eats up more memory when doing similar
> tasks.  Even StarOffice 5 (which was the first version I used) was all
> but unusable execpt on the highest-end Sun workstations that were
> available when it came out.
>
> OpenOffice 2 on my fast PC at work is a lot more sluggish than Office on
> my relatively old Macintosh at home or my even older PC.  For anyone who
> can touch-type above 40wpm and knows his way around the program well
> enough to use hotkeys rather than mouse about aimlessly, those sorts of
> delays are a constant nag.
>
> OS X is OpenStep 7 with some FreeBSD userland utilities and man pages.
> It's not built off FreeBSD any more than FreeBSD is built off GNU/Hurd.
>
> And it's not like I don't use open-source software constantly (I'm on a
> FreeBSD system right now).  I just snicker at the Linux prophets who
> scream "Linux will be mainstream this year" while they still have no
> clue why end users -don't- run Linux.  It isn't about how shiny you can
> make the desktop or how many window managers you can have or how many
> different ways you can animate a window iconifying itself.  It's about
> whether stuff works out of the box, with minimal fuss, and at an
> acceptable speed.
>
> My experience has been that OpenOffice and the recent releases of the
> RPM-based Linux distributions miss that point entirely.
>
> 69473 jp    5  20    0    99M 71484K kserel   0:04  0.00% soffice.bin
>
> That's OpenOffice with one empty document open.
>
> 2674 Microsoft  1.8%  0:02.79  2 80   441  17.3M  57.2M  34.4M 444M
>
> That's Word 2004 for Macintosh with a long, complicated legal document
> open.  It's nearly half the size!
>
> Here's Firefox with nothing open:
>
> 69921 jp   5  20    0 48600K 36656K kserel 0   0:02  0.00% firefox-bin
>
> Here's Safari with nothing open:
>
> 2685 Safari 0.0%  0:01.52   5   113   182  4.62M  21.5M  12.7M 357M

Khtml is *simple* compared to gecko. Actually, I don't get your point
there. You're comparing a partly closed source browser for the Mac
with an open source multi-platform browser based on a different
rendering engine tp prove what?

> I recall the numbers for WinWord and Internet Explorer being similar,
> but I don't have the bandwidth to rdc to a box and find out.
>
> At that rate, I have to buy twice as much computer to use the "free"
> software than I do to run the "non-free" software!  I'm willing to use
> FireFox over IE when I'm on a Windows box just because IE is such a
> flaming ball of garbage, but it always leaves me with a sinking feeling
> of "-This- is the best we could do, over a decade after the web went
> mainstream?"

And only a few years after the end of a rather ugly browser war and
subsequent monopolisation of the web browser market by an utter piece
of crap? Given that there are now several high-quality browsers based
on two good-to-excellent rendering engines (khtml and Gecko) that is
not a bad result at all. When I first installed Mandrake (five years
ago) the software that came closest to being usable was netscape 4.
YUCK!

> > I'm not sure how the Apple support model works, but with Microsoft,
> > they won't talk to you on the phone without a credit card number.
>
> I don't so much care.  My usage of Office is mainly to read other
> peoples Office documents and to print envelopes.  Well, I use Excel as
> well, primarily because there are no free software tools that fill the
> same gap nearly as well.  I haven't found so many bugs in Office for the
> limited things that I do.

The good thing with free software is that even if you decide support
is not worth it, you can still get (ymmv) help from the community.

> > I think people overestimate the importantance/usefulness of commercial
> > support in deciding if they should go with free or commercial
> > software.
>
> Absolutely.

Probably. I'm in no position to argue about that :)

> But the sorts of things that are broken in the typical Open-source
> equivalent program are much more fundamental than the things that were
> broken in the original payware program.  This is understandable, since
> the clone program hasn't been out nearly as long and typically involves
> end-users much earlier in the development process than the corresponding
> commercial program would, but nont of that matters to an end user who
> just wants things to work.
>
> > (Unless you're a big customer, and you're deciding if you want to
> > renew your $500k/year support contract, of course -- then we'll fix it
> > right away.)
>
> Not always even then, as $ork has discovered with a particular EDA tool.
>
> I guess I should boil it down:  When I want to hack code (either my own
> or, lately, the FreeBSD kernel source), I don't really mind when things
> aren't just-so; if they were, I wouldn't have anything to do.  When I
> want to do budget calculations, print address labels, or do other
> mundane computer tasks, I want the box to be as unobtrusive and fast as
> possible.  FOSS only really "gets it" with regards to developer tools
> (like vim, which is just perfect).

That's a chronic problem. The model just works best for those project
where you can expect greater interest from the users. The upside is
that a LOT more people use (let's say) Firefox or OpenOffice.org then
Vim.

> But vim won't put Linux on Grandma's PC.

Totally true.

--
Frank Van Damme