Tcl/TK is indeed dying
in [TCL]
|
Prev: www.prevayler.org anyone do an in ram data base with transactionlog in tcl?
Next: accessing GUI with TWAPI (was: Accessing GUI of CAD through Tcl/Tk)
From: Uwe Klein on 16 Dec 2009 04:52 Rodericus wrote: > On 16 Dez., 00:50, Bruce Hartweg <doNOTmai...(a)example.com> wrote: > > >>object orientation in the core has zero impact to anything you >>want to do with tcl. > > > A fat Tcl interpreter has impact to all wanting to see Tcl linked to > other software and embedded in hardware, because fatness make it less > attractive to those that do not like fatness and do not like to link a > fat interpreter as tool command language to their meager software they > write. Those liking fatness will not have a problem linking a meager > interpreter. The solution is to clear distinguish what is the core of > Tcl (and also Tk) and what is an extension that may be make fat to > satisfy those that like fatness. Is it realy so difficult to > understand? > > The proposal of Pat Thoyts to Arndt Roger Schneider also confirm that > the first do not get the point: Forking Tcl/Tk would only make it > unattractive to potential linkers. Tile as a separated package, as a > complement to a meager Tk, should not be a problem. > > Rodrigo. have you looked at the footprint of (g)libc recently? _That_ is bloat man! On the third hand: nobody beyond the starpackers does static linking anymore. my wish binary ( as an example of an app with embedded tcl _and_ tk ;-) is 6kB small. On small custom (embedded) system I have used jim ( incl. hardware IO and userspace interrupts ) with good results. common to all my tcl apps is that I don't embed an interpreter in a spaghetti monster of c code but provide extra functionality as loadable packages ( tcl or jim ). My apps are interpreted script. OK, I usually arrange it such that I get paid for my work. No need to jump through hoop to hold users up for ransom ;-) Which obviates the need for "hiding" my IP. uwe
From: MSEdit on 16 Dec 2009 09:06 I am not sure where all this bloat you are seeing comes from. My 8.6 full GUI binary is 90k bigger than my 8.5.7 binary for me 90k is not bloat. Martyn
From: Bruce Hartweg on 16 Dec 2009 09:26 Rodericus wrote: > On 16 Dez., 00:50, Bruce Hartweg <doNOTmai...(a)example.com> wrote: > >> object orientation in the core has zero impact to anything you >> want to do with tcl. > > A fat Tcl interpreter has impact to all wanting to see Tcl linked to > other software and embedded in hardware, because fatness make it less > attractive to those that do not like fatness and do not like to link a > fat interpreter as tool command language to their meager software they > write. Those liking fatness will not have a problem linking a meager > interpreter. The solution is to clear distinguish what is the core of > Tcl (and also Tk) and what is an extension that may be make fat to > satisfy those that like fatness. Is it realy so difficult to > understand? > What you don't get is your arbitrarily definition of "Fatness" you want things you want, but not other stuff. You are running this link with databases, webserservers, image libraries etc. so the physical size of the Tcl interpreter is not *really* an issue. if you were doing embedded systems and said I only have x bytes and Tcl is now Y bytes (where Y > x) then you would have a gripe but there is nothing actually preventing you from using it other than your preconceived notions and attitude. Bruce
From: Rodericus on 16 Dec 2009 17:04 On 16 Dez., 15:26, Bruce Hartweg <doNOTmai...(a)example.com> wrote: > What you don't get is your arbitrarily definition of "Fatness" > you want things you want, but not other stuff. I have just done man perl: >>The Perl motto is "Theres more than one way to do it."<< Perl is always there when I install Free- or OpenBSD, Tcl/Tk is unfortunately not anymore in the base distribution of FreeBSD. Perhaps because it is trying to compete with perl. You have now at least 15 ways: (1) Like LISP (but with less parenthesis). (2) Typical Tcl, writing scripts that write and change themselves before running over their change. (3) like C. (4) Object oriented. (5) - (15) Every combination of some of the above possibilities. Having a goto command, you could also have the possibility of writing like FORTRAN and the combinations, a total of 31 possibilities. You can argue that this is good, you have the freedom to chose a style and ignore the other possibilities, and this is correct as far as you are the only one that write and read what he writes. Years ago there were big discussions for and against the goto statement (that I like very much). In C you have the goto, in Tcl not, perhaps because it was considered bad style. The question of style has an impact in software developement, and the restriction in the language a sense. I like the description of Tcl as a "Toy Language", it was only a misunderstanding of its minimality and simplicity that allowed to have a clear view over the whole language, that made it easy to learn it completely. This was an advantage that seems to be not more understood. I am still using Tcl8.4, I was disgusted as I saw widget repetitions and the object orientation. If Tcl/Tk is not jet fat, then it is perhaps on the way. I think, obeject orientation in a scripting language is fat. I think object orientation self is fat hiding algorithms, but this is an opinion. Rodrigo.
From: Georgios Petasis on 16 Dec 2009 18:20
O/H Rodericus έγραψε: > On 16 Dez., 15:26, Bruce Hartweg <doNOTmai...(a)example.com> wrote: > >> What you don't get is your arbitrarily definition of "Fatness" >> you want things you want, but not other stuff. > > I have just done man perl: >>The Perl motto is "There's more than one > way to do it."<< > > Perl is always there when I install Free- or OpenBSD, Tcl/Tk is > unfortunately not anymore in the base distribution of FreeBSD. Perhaps > because it is trying to compete with perl. You have now at least 15 > ways: > > (1) Like LISP (but with less parenthesis). > > (2) Typical Tcl, writing scripts that write and change themselves > before running over their change. > > (3) like C. > > (4) Object oriented. > > (5) - (15) Every combination of some of the above possibilities. > > Having a goto command, you could also have the possibility of writing > like FORTRAN and the combinations, a total of 31 possibilities. You > can argue that this is good, you have the freedom to chose a style and > ignore the other possibilities, and this is correct as far as you are > the only one that write and read what he writes. Years ago there were > big discussions for and against the goto statement (that I like very > much). In C you have the goto, in Tcl not, perhaps because it was > considered bad style. The question of style has an impact in software > developement, and the restriction in the language a sense. I like the > description of Tcl as a "Toy Language", it was only a misunderstanding > of its minimality and simplicity that allowed to have a clear view > over the whole language, that made it easy to learn it completely. > This was an advantage that seems to be not more understood. > > I am still using Tcl8.4, I was disgusted as I saw widget repetitions > and the object orientation. If Tcl/Tk is not jet fat, then it is > perhaps on the way. I think, obeject orientation in a scripting > language is fat. I think object orientation self is fat hiding > algorithms, but this is an opinion. > > Rodrigo. > I didn't understand the relation with perl. Perl also has OO. And the difference in sizes of the involved dlls is about 60K. Do you compare with perl because perl is a "slim" language, or you are considering perl also as fat? George |