http://slashdot.org/firehose.pl
Keith.
Well, somebody had to do it :-) The announcement of Perl's new release was
there on the 19th (two days ago)...
Vote for it:
http://slashdot.org/firehose.pl?op=view&id=435816
Yours,
-mi
That summary is far too short, to be totally honest. It's around 20
words when the ideal slashdot summary is in the 60-100 word range.
Ideally, it should reference not just the download site but also the
article on OS News too.
http://www.osnews.com/comment.php?news_id=19073
Donal.
The new post is much better, but is still ranked quite low. It isn't
going to get to the main page unless some more people vote it up.
if newness counts, mod this one up;
http://slashdot.org/firehose.pl?op=view&id=437224
It seems like the Slashdot community has repeatedly made it clear they
don't care about Tcl. My prediction is that one article gets posted,
followed by 80,000 retarded quasi-programmers bashing Tcl, and drowning
out the well informed people that do use Tcl.
And I wish it was just slashdot... Search for Tcl in this discussion:
http://lambda-the-ultimate.org/node/908
Neil Madden and David Welton tried to defend Tcl, only to be drowned out
IMO.
One idea in the Tcl chat last night was to not put emphasis on how easy
Tcl is to use, because realistically it isn't. Being a *good* Tcl
programer isn't easy, especially if you come from a C background, and
assume [set set set] is impossible.
Here's how to react to a typical flame:
idiot: "Tcl sucks, it has quoting issues."
programmer: "Tcl is not for dummies -- like yourself. Perhaps you
should try Java, Python, or Perl.
Or this lovely situation:
idiot: "Tcl is slow and it lacks ..."
programmer: "More likely you are slow. Tcl is fast for people that
understand it. Perhaps you should try a lesser language like BASIC."
George
--
Our heart glows, and secret unrest gnaws at the root of our being.
Dealing with the unconscious has become a question of life for us.
-- Carl Jung
It took a bit longer (and a more carefully written article) but there's
now an announcement story out:
http://it.slashdot.org/article.pl?sid=07/12/22/2315246
As a lesson for people in the future, if you want a story to get picked
up by slashdot, you *must* take care to write it well. They virtually
never put up stories that are badly written so you've got to write the
story you want to read (and preferably link to several other sites that
also discuss the issues...)
Donal.
In article <zqvbj.41236$036....@fe1.news.blueyonder.co.uk>,
No problem. I just wrote what I wanted to read. :-)
Donal.
Next thing you know, Donal will be the resident marketing expert in
the Tcl Core Team:-)
Shhh, don't scare him off! : )
> It seems like the Slashdot community has repeatedly made it clear they
> don't care about Tcl. My prediction is that one article gets posted,
> followed by 80,000 retarded quasi-programmers bashing Tcl, and drowning
> out the well informed people that do use Tcl.
>
> And I wish it was just slashdot... Search for Tcl in this discussion:
> http://lambda-the-ultimate.org/node/908
Wo cares? It's a small fraction of people who have enough time to write
and read such bashings. I wouldn't see this too negative.
I would rather love to see people emphasizing the real strengths of Tcl
as a language. E.g. the choice of OO systems and how this relates to the
-on "lambda-the-ultimate" obviously beloved- Meta classes
(http://lambda-the-ultimate.org/node/749). Do they know XOTcl, and that
this was around, long before the paper submission and the IBM article
that are mentioned on this site?
Or the possibility to entirely redefine the language and add constructs
as needed.. is there any other language in the world where you can do
this in the same fashion?
> One idea in the Tcl chat last night was to not put emphasis on how easy
> Tcl is to use, because realistically it isn't. Being a *good* Tcl
> programer isn't easy, especially if you come from a C background, and
> assume [set set set] is impossible.
IMO the main problem for Tcl's popularity is it's syntax. Most people
don't like prefix languages, they don't know how to handle them (BTW, I
loved the syntax from the first moment). The prefix notation requires a
different thinking - but it also makes features possible that don't
exist anyhere else. Especially in conjunction with the "everything is a
string" paradigm this is extremely powerful.
Ok, most programmers don't care about such features as language
extension and redefinition. But is it necessary to count on those people?
>
> Here's how to react to a typical flame:
Here is how I react: if I feel that the orig poster is ignorant, I don't
answer at all, and I don't care. If I feel that he is kindof interested,
I explain the strength's in as few understandable words as possible. The
latter is mostly the case in oral conversations...
Eckhard
> Ok, most programmers don't care about such features as language
> extension and redefinition. But is it necessary to count on those people?
I mean foremost: is it necessary to count on people who bash Tcl because
of it's syntax...? It is certainly right to count on people who not want
to redefine the programming language and just use it as it is ;-).
Eckhard
[snip]
> I would rather love to see people emphasizing the real strengths of
> Tcl as a language.
I love objective, unbiased, non-religious analysis. :-)
> E.g. the choice of OO systems and how this relates
> to the -on "lambda-the-ultimate" obviously beloved- Meta classes
> (http://lambda-the-ultimate.org/node/749). Do they know XOTcl, and
> that this was around, long before the paper submission and the IBM
> article that are mentioned on this site?
> Or the possibility to entirely redefine the language and add
> constructs as needed.. is there any other language in the world where
> you can do this in the same fashion?
Without knowing what do you mean with "same fashion", well, I'll dare to
say Common Lisp.
>> One idea in the Tcl chat last night was to not put emphasis on how
>> easy Tcl is to use, because realistically it isn't. Being a *good*
>> Tcl programer isn't easy, especially if you come from a C
>> background, and assume [set set set] is impossible.
I totally agree with this!
> IMO the main problem for Tcl's popularity is it's syntax. Most people
> don't like prefix languages, they don't know how to handle them (BTW,
> I loved the syntax from the first moment). The prefix notation
> requires a different thinking
Most people has no problem writing [foo bla bla] instead of foo(bla,
bla). What they hate is having to write [< [+ bla 3] [/ [- x 4] 3]]. But
Tcl has not this problem, unlike Lisp.
> - but it also makes features possible
> that don't exist anyhere else. Especially in conjunction with the
> "everything is a string" paradigm this is extremely powerful.
I'm not sure EIAS is a net gain, but I'm not a Tcl expert either. IMHO,
Lisp does this much better with code-is-data-is-code.
Tcl excels at some things (that's because I use it!): easy to embed,
lightweight, rich library, Tk, "easy" (in the sense that is easy to get
something simple up and running), good community, etc. But I wouldn't
try to sell it as a superior general purpose *language*.
[snip]
--
Oscar
>> I would rather love to see people emphasizing the real strengths of
>> Tcl as a language.
>
> I love objective, unbiased, non-religious analysis. :-)
Definitely :).
> Most people has no problem writing [foo bla bla] instead of foo(bla,
> bla). What they hate is having to write [< [+ bla 3] [/ [- x 4] 3]]. But
> Tcl has not this problem, unlike Lisp.
Well, [< [+ blah 3] [/ [- x 4] 3]] is really bad.. for non-Lisp freaks.
But [set i 0] is already "bad enough", compared to i=0. I think there
are just two ways: either one likes that style or not. Nothing in between.
> I'm not sure EIAS is a net gain, but I'm not a Tcl expert either. IMHO,
> Lisp does this much better with code-is-data-is-code.
I don't know Lisp well enough, but when I understand
"code-is-data-is-code" right from a Tcl'ers perspective, then this
exists as well:
set cmd {
foreach e $l {
do $e
}
}
...
# later
eval $cmd
cmd is data is code is data?
> Tcl excels at some things (that's because I use it!): easy to embed,
> lightweight, rich library, Tk, "easy" (in the sense that is easy to get
> something simple up and running), good community, etc. But I wouldn't
> try to sell it as a superior general purpose *language*.
Does this "superior general purpose" language exist anyway?
Eckhard
[snip]
> Well, [< [+ blah 3] [/ [- x 4] 3]] is really bad.. for non-Lisp freaks.
> But [set i 0] is already "bad enough", compared to i=0. I think there
> are just two ways: either one likes that style or not. Nothing in
> between.
Old BASIC has LET x = 10, which is worse than Tcl (and it doesn't offer
any advantage over x = 10, whereas Tcl's set does). Old BASIC was used
by millions of people without too much complain. It ruined, by means of
mind-constraining, so much actual programmers that their creators should
be prosecuted! (BASIC, COBOL, Fortran are mind-constraining languages
due to their rigidness, opposed to Tcl, Lisp, Forth, etc which are so
malleable that creative people actually enjoy programming).
>> I'm not sure EIAS is a net gain, but I'm not a Tcl expert either. IMHO,
>> Lisp does this much better with code-is-data-is-code.
>
> I don't know Lisp well enough, but when I understand
> "code-is-data-is-code" right from a Tcl'ers perspective, then this
> exists as well:
>
> set cmd {
> foreach e $l {
> do $e
> }
> }
> ...
> # later
> eval $cmd
>
> cmd is data is code is data?
Not as much as in Lisp's case. You have the data-is-code, but you can't
go the code-is-data path as easily as Lisp, which has an even simpler
syntax than Tcl. This allows the existence of a powerful macro system:
functions that inspect parts of the program and take some action, that
usually is spitting more code. But my point was that Lisp's data is not
confined to strings: you have cons cells, for instance, which are a pair
of pointers and hence the basis for complex data structures.
>> Tcl excels at some things (that's because I use it!): easy to embed,
>> lightweight, rich library, Tk, "easy" (in the sense that is easy to get
>> something simple up and running), good community, etc. But I wouldn't
>> try to sell it as a superior general purpose *language*.
>
> Does this "superior general purpose" language exist anyway?
IMHO, yes. And, as is the case with true religions, there are several
:-) We all agree that Tcl is superior to BASIC, don't we? Well, I'm
convinced that it is possible to convince an honest audience that, in
most cases, language X allows solving problems on a more convenient way
than language Y. You can do that for the Tcl > BASIC case, and IMO the
same can be done for the Scheme > Tcl case.
Although, as always, the programmer is the key piece. No doubt there are
languages that attracts certain types of minds better than others (think
Haskell, or the pure functional crowd in general for a clear case). So
the culture, personality, etc of the programmer is crucial on
determining how effective he is solving a problem using a language. This
is something that most programming language proselytisers fail to admit
(unlike George Peter Staplin with his [set set set] above). It is
perfectly possible that you convince someone that language X solves a
problem more efficiently than language Y but *he* is unable to build
solutions on language X as efficiently as he does on Y.
--
Oscar
LISP, and to an even greater extent than LISP and Tcl, FORTH. FORTH is
the only language I know of that lets you define new literal types.
--
Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.