[Openmcl-devel] OpenMCL Versions

[Justin Fagnani-Bell]

You make, of course, some good points about the size of the  
interested audience and the amount of donations needed to fully  
support a IA32 port. I'm surprised at the amount of work you  
estimate. I know hardly anything about the x86 ISA or what's involved  
in a Lisp implementation (I gather it's very low level). It's too bad  
that it can't be written so that a few #DEFs and compiler flags will  
produce either a IA32 or EM64T version.

Would partial finding make a port more worthwhile? Probably not that  
much, but I think you should have a way to donate anyway, since your  
work is still valuable to many, many people who might be interested  
in saying thank you and helping the cause.

For me, I'm looking to LispWorks which is one of the few (only?)  
other Lisps that has Cocoa and Intel support. Unfortunately I can't  
afford the Professional Edition and the free Personal Edition is a  
version behind and so doesn't support Intel yet. In a couple of years  
when I upgrade to a 64-bit Mac I'll hopefully be able to use OpenMCL.

Thanks,
   Justin


On Aug 17, 2006, at 12:32 PM, Gary Byers wrote:

> A port to 32-bit Intel hardware would probably take 3-4 months.   
> (That's
> less time than the x86-64 port took, but some of the infrastructre it
> uses can be shared.)
>
> If you multiply 3.5 by the amount of money needed to support my  
> decadent
> extravagant lifestye each month ... you wind up with a number that's
> in the mid-5-digits in US$.
>
> Let's say that 5 millon 32-bit only Macs have been sold to date.  Of
> that 5 million, some number (probably a fairly small number, sadly)
> want to do lisp development in OpenMCL on their shiny new 32-bit Intel
> Macs.  Of that number, -some- individuals may want and be able to
> kick in money to pay for that port.
>
> Are there 100 people who'd kick in $500 ?  50/$1000 ?  If so, they've
> been mostly silent over the last year (and we haven't aggressively
> tried to pursue this idea.)
>
> Are there institutional users who want to make a significant
> investment in something that doesn't address their long-term needs ?
> That seems to be pretty much the same story (though we're very glad
> to have found funding to do the 64-bit port.)  Admittedly, the
> way that I'm phrasing that makes the assumption that a 32-bit Intel
> port wouldn't address the long-term needs of such a user.
>
> Is it in Clozure's interest to fund OpenMCL development ourselves,
> when we're able to and when it involves something strategically
> important ?  Yes, we've done so and plan to do so when we can.
>
> Do at least a few other things seem to be more "strategically
> important" than a 32-bit port ?  Yes, and people on this list
> generally enumerate them on a regular basis.  That assessment
> might be inaccurate and priorities could change.
>
> I don't like losing users and potential users.  I didn't like Apple's
> decision to switch hardware platforms twice over a two-year period
> (ppc->ia32, ia32->x86-64) or to leave Rosetta incomplete, either.
>
> If there are institutional users that we haven't been in contact with
> or 500 to 1000 seriously interested individuals, I will be very
> (pleasantly) surprised.
>
> On Thu, 17 Aug 2006, Justin Fagnani-Bell wrote:
>
>>
>> On Aug 16, 2006, at 6:21 PM, Gary Byers wrote:
>>
>>> I think that there are stories on the net
>>> indicating that people have successfully done CPU upgrades of  
>>> some Mac
>>> Intel systems, and it would be nice if this was a supported option
>>> for people who want a speed boost and 64-bit capability but don't
>>> want to replace their entire brand-new systems.
>>
>> minis and iMacs are socketed, and there are reports of Merom working
>> in both. There's still going to be millions of them out there running
>> 32-bit chips, and of course, millions of MacBook and MacBook Pro
>> owners who can't upgrade.
>>
>> -Justin
>> _______________________________________________
>> Openmcl-devel mailing list
>> Openmcl-devel at clozure.com
>> http://clozure.com/mailman/listinfo/openmcl-devel
>>
>>