Movatterモバイル変換
[0]ホーム
Musing out loud... [Why not Smalltalk?]
Piercarlo Grandipg_nh at sabi.Clara.co.UK
Tue Apr 17 20:17:57 EDT 2001
>>> On Sun, 15 Apr 2001 09:09:13 -0400, "Eric Clayberg">>> <clayberg at instantiations.com> said:[ ... ]>>>> Javaites probably know that the Self VM project was where Sun's>>>> 'hotspot' technology was developed.>> Here's how Sun presents it>>http://java.sun.com/products/hotspot/whitepaper.html > "The Java HotSpot performance engine architecture addresses the Java > programming language performance issues described above by using > adaptive optimization technology. Adaptive optimization is the fruit > of many years of research into object-oriented language > implementation performed by the Self group at Sun's research > division.clayberg> That's a pretty funny example of revisionist history for thoseclayberg> of us who were around when Anamorphic first built HotSpot (forclayberg> Smalltalk) and shopped it around (in 1996). [ ... ] What Sunclayberg> says above is sort of true, but leaves out quite a fewclayberg> intermediate steps. Several of the ideas for HotSpotclayberg> originated with Self (or were well known in the literature),clayberg> but it took the Anamorphic team to pull them all together andclayberg> prove that they worked in concert. The Self group did notclayberg> create HotSpot; Sun bought it off the shelf.Uhhhhhhmmmmmm, yes and no. It looks like that Hotspot the Anamorphicproof of concept was bought off the shelf for a rather large price; butSun's "adaptive optimization technology", whether called HotSpot or not,was definitely developed by the Self group and collaborators at Sun Labsand related organizations:http://WWW.Sun.com/research/self/papers/papers.htmlThis development succeeded to the point that Mario Wolczko's fairlystraighforward Smalltalk-80 implementation on top of Self:http://WWW.Sun.com/research/self/release/smalltalk.htmlhttp://WWW.Sun.com/research/self/release/Self-4.0/manuals/smalltalk.ps.Zwas often much faster than any then existing Smalltalk-80implementation, even popular commercial ones.Thus I find it a bit bizarre to imagine that "it took the Anamorphicteam to pull them all together and prove that they worked in concert",because the Self group not only largely developed them, but they didthat themselves (actually it was mostly just one guy for theSmalltalk-80 clone) and in the most straightforward way possible.
More information about the Python-listmailing list
[8]ページ先頭