First proposal: what should LispOS feel like?
Wed, 30 Apr 1997 14:32:13 +0200
Seems like we have found a third area to start digging in ..
>>"KM" == Kelly Murray wrote on Tue, 29 Apr 1997 17:40:17 -0700:
>> What would motivate anyone but LISP users to adopt such a system?
KM> Only Lisp programmers will use this Lisp (Silk) Machine. Our
KM> goal is to create/convert people into Lisp programmers!
KM> Why would they do that? Because they want/need to develop
KM> maintainable web-based applications quickly using the VirtualLM.
KM> The best platform to run this on is a RealLM which is fast and
KM> efficient, and further gets rid of all that UNIX junk too. We
KM> can dream about RealLM's eventually replacing UNIX,NT, but don't
KM> count on it, and we don't.
So, this sounds like the third area: develop an API that _is used_ by
an (probably award-winning) app like Kelly's Web-App (whatever it
might ultimately result in). Let's have some persistent
object-API (you name it) that is provided by the VirtualLM (or the
RealLM underneath it) and some fine GUI abstraction layer (Free CLIM
?). Start three working groups, which all work tightly together to
form a powerful environment. One works on the RealLM (that is, bring
Lisp on the hardware, probably on top of Linux), one on the VirtualLM,
one on a powerful app and the needed API with the other two groups
pumping support and functionality into this API.
KM> Let me also clarify how I see this working in this regard. I
KM> don't see a single address space machine, though that might be
KM> find for starters. I see a machine where processes have their
KM> own private, transient memory, BUT they share persistent objects
KM> which reside in databases.
Define an API for the stuff, let every (SILK-)application use it. Let
the RealLM/VirtualLM-people implement the necessary stuff.
I believe that the most important thing is exactly this: a
common and shining environment, which is truely implemented and truly
used. Let's have some cake and eat it, too.