Re: [Fresco-devel] Re: ACE

On Sun, Nov 24, 2002 at 10:16:07PM -0700, M. Evans wrote:
> 
> > We're aware of ACE, in fact, and some work has been done to let us
> > work with TAO (The ACE ORB).  The library is a quite heavy, though
> > (the problem with using TAO is that the generated libraries are
> > _huge_), has long compile times, and, aside from TAO, does the same
> > sort of stuff that we have Prague to do.  ACE does _more_ of that
> > sort of stuff, but we don't need more, and Prague is much lighter
> > weight.
> 
> > So while we've definitely learned from ACE, it's just overkill for
> > everything we've needed so far.
> 
> Nathaniel I can't tell you how many times I've heard this lame excuse
> for reinventing wheels -- the overkill excuse.  I never liked it.
> To me, it's always easier to strip things down than to re-invent
> them from scratch.  No offense, that's just my attitude.

You suggest we fork ACE and throw out all the parts we don't need, to
make it slim enough to be usable?  That's absurd.  It is, in fact,
_not_ easier to do that, than to just write the parts we need.  ACE is
huge, complicated, with a _lot_ of internal dependencies.

Saying "it's always easier to strip things down" is simply not true in
the real world, and makes me wonder how much time you've spent doing
actual engineering.

> What this shows is a case where the whole CORBA idea falls apart.
> It doesn't necessarily show that ACE is overkill.  Here we have two
> CORBA systems but you're saying is that it's just too much nightmare
> making them work together.  This kind of stuff is why Microsoft
> has migrated from COM to .NET, which removes all the IDL junk.

This has nothing to do with what I said, and is incoherent.

ACE is not a CORBA system.  I don't see how ACE being overkill as a
POSIX wrapper makes the "whole CORBA idea" untenable.

There is, however, an ORB associated with ACE, called TAO.  It is no
nightmare at all getting TAO and omniORB to work together -- this is
an advantage of CORBA.  I never said that it was a nightmare.

In fact, our code compiles without changes, and (we hope) works
identically, with both TAO and omniORB.  This is another advantage of
CORBA.

However, because TAO uses ACE, it generates extremely bloated
libraries, so huge that Stefan never even tried using them.  This
could be considered a bug in TAO.

And then you pull out this wildly irrelevant thing about Microsoft,
and something about IDL that makes no sense.  .NET is a wildly
different architecture from CORBA; both involve OO programming, and
distributed objects, but with very different approaches.  CORBA is a
communication protocol; .NET is a virtual machine.  It's like
comparing TCP/IP and the Pentium.  And somehow this switch supposed to
be in response to the problem of incompatible implementations, when
AFAIK, COM and .NET have each only ever had one implementation, and
that controlled by MS...

(Yes, I know about Mono, and there have probably been COM workalikes
the the like.  But nothing relevant to MS's decision to switch.)

> Open-source projects are just way too undermanned for me to ever
> accept the overkill excuse.  Build problems are nothing next to
> re-inventing wheels.

This might make sense, if it was in fact more work to do the things
you say.  But it isn't.

-- Nathaniel

-- 
Sentience can be such a burden.

This email may be read aloud.