"Expert" testimony (was: Odd article on Scriptics at SunWorld)
Cameron Laird
James Ingham <jin...@leda.cygnus.com> wrote:
.
.
.
>An issue of more concern is that the Gartner Group expert that the
>reporter contacted said that a main concern was that Tcl was still
>"interpreted" which could cause speed problems, though she had no
>hard numbers. This validates another part of the article, which is
>that Scriptics has some work to do still on the education front, both
>in teaching people what an "Extensible" architecture means (don't do
>your time critical stuff in Tcl...), and about the 8.0 Byte compiler.
>After all, with the ability to write out the byte codes in TclPro, Tcl
>is no more interpreted that Java...
.
.
.
Can someone explain to me the function of nominally
analytic agencies like The Gartner Group? I've
turned *very* negative on them in the last year.
Of course, since all I know about them over that
span is what they're quoted as saying, maybe it's
just transmission noise that makes them all sound
so, well, other-than-value-adding.
Briefly, this is my observation: I can find no
definite relation between what these industry ex-
perts say, and reality. They don't seem to review
work they've already done; they don't seem to know
their subjects from their own experience; and they
simply don't seem to get the point of Science.
--
Cameron Laird http://starbase.neosoft.com/~claird/home.html
cla...@NeoSoft.com +1 281 996 8546 FAX
Mark Hecht
they will invariably make.
...mark
On 2 Sep 1998 21:12:13 -0500, cla...@Starbase.NeoSoft.COM (Cameron
Laird) wrote:
... snip
Alexandre Ferrieux
>
> Briefly, this is my observation: I can find no
> definite relation between what these industry ex-
> perts say, and reality. They don't seem to review
> work they've already done; they don't seem to know
> their subjects from their own experience; and they
> simply don't seem to get the point of Science.
If any flame war comes out of this very last sentence, count me on your
side.
I'll bring ammunitions and sandwiches.
Now I can't help kicking back on the technical point too.
Basically, this is the ages-old 'Tcl is slower-than-X' story.
Heck ! My blood's approaching a phase transition (guess which one ?)
whenever I hear that.
So again:
(Cameron, I know you know, I'm just perusing the NNTP broadcast channel
:)
The emerging Good Thing about software reuse is 2-layer blackboxing: a
higher layer (highly readable shell) achieves human efficiency, a lower
layer (thoroughly optimized binary black-box components) achieves CPU
efficiency. No intelligent person would try to achieve both with the
same language (right-tool-for-the-right-job!).
(People interested could use
ftp://ftp.neosoft.com/pub/tcl/sorted/info/doc/tclarch.txt)
Hence *any* speed comparison with a single-level language like Java is
pointless. Tcl wouldn't even *try* to do an FFT quickly. Java does try.
And imperfectly, since a JVM will always be slower than the supporting
hardware. Worse, Java is simply not credible in the higher-layer job. So
it's bad for both :)
Okay so now please folks, throw sharp and heavy things here -> [metal]
[glass] [stone]
(Sorry Cameron I really couldn't help :)
-Alex
Cameron Laird
Alexandre Ferrieux <alexandre...@cnet.francetelecom.fr> wrote:
>Cameron Laird wrote:
.
.
.
>If any flame war comes out of this very last sentence, count me on your
>side.
>I'll bring ammunitions and sandwiches.
[comments on software
design in a "componentized"
milieu]
.
.
Delighted.
Shaw has the (anti-)hero of *Arms and the Man* explain that
wisdom for an infantryman is to pack chocolates in his knap-
sack, for they're always at least as useful as the bullets
and grenades that are supposed to belong there. That's a
rough paraphrase; it's been at least fifteen years since I
read the play.
In any case, Shaw's style is confident and distinctive enough
that he always sounds most like himself. I've started to
catch on how expert he in fact was at re-using the work of
others (that's where the programming content of this thread
was headed, incidentally). Is there an obvious source for
the literary conceit about candies in the cartridge case from
which Shaw lifted his passage?
Stefaan A Eeckels
Alexandre Ferrieux <alexandre...@cnet.francetelecom.fr> writes:
> The emerging Good Thing about software reuse is 2-layer blackboxing: a
> higher layer (highly readable shell) achieves human efficiency, a lower
> layer (thoroughly optimized binary black-box components) achieves CPU
> efficiency. No intelligent person would try to achieve both with the
> same language (right-tool-for-the-right-job!).
>
> (People interested could use
> ftp://ftp.neosoft.com/pub/tcl/sorted/info/doc/tclarch.txt)
>
> Hence *any* speed comparison with a single-level language like Java is
> pointless. Tcl wouldn't even *try* to do an FFT quickly. Java does try.
> And imperfectly, since a JVM will always be slower than the supporting
> hardware. Worse, Java is simply not credible in the higher-layer job. So
> it's bad for both :)
>
> Okay so now please folks, throw sharp and heavy things here -> [metal]
> [glass] [stone]
> (Sorry Cameron I really couldn't help :)
IMHO (as always :-) you're spot on. Tcl/Tk brings the UNIX spirit to
GUI applications (have programs that do one thing, and do it well).
Take care,
--
Stefaan
--
PGP key available from PGP key servers (http://www.pgp.net/pgpnet/)
___________________________________________________________________
Perfection is reached, not when there is no longer anything to add,
but when there is no longer anything to take away. -- Saint-Exupéry
Kurt Wall
% Cameron Laird wrote:
% >
%
[snip]
% Okay so now please folks, throw sharp and heavy things here -> [metal]
% [glass] [stone]
Isn't it paper, rock, scissors?
% (Sorry Cameron I really couldn't help :)
%
% -Alex
Kurt