Why Clang

Chad Perrin perrin at apotheon.com
Fri Jun 22 23:58:21 UTC 2012

On Fri, Jun 22, 2012 at 08:28:17AM +0200, Wojciech Puchar wrote:
> >
> >biggest problem with what you propose, though, is that it would destroy
> >the social factors in development of the FreeBSD system that make it what
> >it is, and thus destroy FreeBSD itself, as far as I am concerned.
> I am not sure, as long as clients would be treated seriously!

I look at large corporate software vendors and see them treating
customers seriously maybe 2% of the time at best.  In this case, most of
the developers and project managers of FreeBSD are also "customers",
which changes things significantly.

> >
> >I would have thought that even you should be able to understand that
> >without help.
> another personal attack? I though i talk with adults.

1. It's a comment on your tendency to ignore substantive arguments from
other people, including probably half a dozen (so far) lengthy
explanations of factors you refuse to consider written by *me*.

2. You're a hypocrite, pretending you're an innocent victim of personal
attacks, given the way you go around making personal attacks on everyone
else with a broad brush.  I've commented on that, too, but -- like much
of the rest of what I've said -- you simply ignored it.

> >
> >Turning it into a commercial enterprise rather than an open source
> >project would probably turn it into a project that is driven about 60% by
> >corporate politics and 40% by marketing BS, with no room left over for
> >quality except as needed to support the minimum credibility its CEO deems
> >necessary to support those two concerns.
> It depends solely on development team.

I take it you don't know anything at all about how public corporations
manage their development teams.  That, or you're being disingenuous.

It depends on the development team, and the priorities they choose to
pursue first, right now.  Under the stewardship of a publicly traded
corporation, it would depend on the CEO, the board of directors,
marketing, PR, and the accounting department, and the priorities *they*
choose to pursue first, instead.

> For now - as we see - it's decision are driven by money.
> But not all users money but few selected large users.

It's not *just* a decision driven by money.  Money applies, certainly,
but not as much as it would if FreeBSD were a for-profit public
corporation rather than a community-driven open source project.  When you
say this, by the way, you ignore something like 90% of the perfectly
reasonable additional motivating factors that have been brought up.  I
suppose I should not expect any different by now, given the strong track
record you've managed to establish just in this one extended discussion.

> >
> >"Worse" based on a couple of very narrowly applicable metrics derived
> There will be IMHO soon good compiler available. it's highly
> probable that pcc would improve a lot, for now it is small, quick
> but doesn't produce good code for new CPUs. But it probably will
> improve.
> CLANG is already great bloat, and will be worse.

Binary size and minuscule benchmark variations are all you see.  It is
ludicrous to watch you close your eyes, stick your fingers in your ears,
and shout "lalalalalalalala" so consistently to prevent any other factors
involved in compiler choice from entering your mind -- such as good
output from a compiler that will be stable and do what you expect.

> No amount of money will fix it, actually too much money will hurt.

. . . and yet you want to turn the FreeBSD project over to Microsoft (or
the equivalent).  You contradict yourself.

Chad Perrin [ original content licensed OWL: http://owl.apotheon.org ]

More information about the freebsd-questions mailing list