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: Rodericus on 13 Dec 2009 07:24 Tk/Tk was my predilect language for small programms, because it was minimalistic and expresive, low weight and extensible, lisp and C similar, ideal for embedding it in other programms. Now it is getting fat and "object oriented" with a lot of unnecesary "features" that would belong to extensions for special purpose applications. It is getting a "Cool Programming Language (CPL)" for cool people, not any more a "Tool Command Language". I think, a splitting and a renaming of the cool language to something like Cpl/Tk#++ would have been a much better approach. I think this is the result of having very good developers not knowing what to do. Please, dont consider this posting a flame war provocation: it is my oppinion. Rodrigo Readi
From: APN on 13 Dec 2009 10:33 To borrow an old slogan from oldsmobile, this indeed is "Not Your Father's Tcl" (nyf-T, how's that for a new language name - nifty, no?). But that is a good thing. Personally, I feel many/most of the new features greatly expand the scope and size of software systems you can build using Tcl. So I suppose that means it is no longer restricted (if it ever was) to be solely viewed as "tool" language. In that sense, I understand your point of view but if you feel a language should not "grow" features, what's preventing you from sticking with older, smaller versions (even Tcl 7.6) for embedding? /Ashok On Dec 13, 5:24 pm, Rodericus <sc...(a)web.de> wrote: > Tk/Tk was my predilect language for small programms, because it was > minimalistic and expresive, low weight and extensible, lisp and C > similar, ideal for embedding it in other programms. Now it is getting > fat and "object oriented" with a lot of unnecesary "features" that > would belong to extensions for special purpose applications. It is > getting a "Cool Programming Language (CPL)" for cool people, not any > more a "Tool Command Language". I think, a splitting and a renaming of > the cool language to something like Cpl/Tk#++ would have been a much > better approach. I think this is the result of having very good > developers not knowing what to do. Please, dont consider this posting > a flame war provocation: it is my oppinion. > > Rodrigo Readi
From: Kevin Kenny on 13 Dec 2009 11:11 Rodericus wrote: > Tk/Tk was my predilect language for small programms, because it was > minimalistic and expresive, low weight and extensible, lisp and C > similar, ideal for embedding it in other programms. Now it is getting > fat and "object oriented" with a lot of unnecesary "features" that > would belong to extensions for special purpose applications. It is > getting a "Cool Programming Language (CPL)" for cool people, not any > more a "Tool Command Language". I think, a splitting and a renaming of > the cool language to something like Cpl/Tk#++ would have been a much > better approach. I think this is the result of having very good > developers not knowing what to do. Please, dont consider this posting > a flame war provocation: it is my oppinion. It's hard to ignite a flame war here. :) I'm sorry that recent developments make the language fail to serve your needs. Alas, you really have not given me much to go on to address your criticism. Most such complaints fall into two major categories. Either the complaint is that the image has become too large to fit comfortably on smaller machines, or that the language has become too difficult to learn. Regarding the first complaint (that the language doesn't fit any more): Many of us have had the experience that even "small" platforms have grown over time, Moore's Law being what it is. A couple of megabytes of object code would have been quite a large program when Tcl was introduced, but now represent a smallish library that can be included almost without thought on anything bigger than a handheld platform. Over the last decade or so, there have been several attempts to get a group of people to evolve the language from a base of Tcl 6.7 or Tcl 7.6 so that a tiny version is available for those who live with such tight constraints. They haven't been very successful. My guess is that it represents a vanishingly small niche: small platforms for which the vendor offers user-programmability and that aren't capable enough for a newer version. That immediately rules out most mobile devices. Mobile vendors view applications as a profit centre, and typically forbid user programming; the iPhone application store, widely castigated by developers, is actually nearly the most open model out there! Moreover, small platforms that are not tied to the telephone system tend to be deeply embedded in devices. Those that have the resouces to develop for such beasts also have the resources to pull a smaller Tcl over from SourceForge and adapt it to there needs. Since older versions do get downloaded occasionally, I presume there are still developers doing just that. If the complaint is rather that the abundance of new features has made the language too hard to learn, I have sympathy, but not much to offer. Surely freezing the language and not offering new development would also be a "death strategy;" the choice of what goes in and what stays out is quite a narrow tightrope to walk. Object orientation, to use your example, is something that has been extremely widely demanded, and for which at least half a dozen mutually incompatible extensions had been in wide use. All of the extensions suffered from performance (and usually stability) issues, caused in large measure by the failure of the core language to export appropriate interfaces for them to use. It seemed to me that the existence of multiple extensions for the same task, and the immense difficulty of getting those extensions to perform adequately, was a compelling argument that we had here a missing core language feature. Moreover, OO as specified in TIP 257 does not (thankfully!) make Tcl into an aggressively "object oriented" language like Java or Python. A programmer can ignore the '::oo' namespace entirely and still contrive to lead a useful life without it. Reading between the lines, I suspect that your posting may also be suggesting that you feel that we have neglected something more important, in favour of implementing features that offer you little benefit. In that case, it would be more effective for you to lobby affirmatively for the development you need. But you have to recognize that volunteers, as we are, have a complex set of motivations. An all-volunteer initiative is somewhat unpredictable in its direction. The compensating advantage is that development is open. You can help steer things, either by contributing your own time and effort, or by paying someone else to do so. And there are developers out there who would be happy to take on core-improvement efforts for pay. Alas, the good ones are expensive, but isn't that always the way of it? In short, you can't please everyone. If your priorities are enough different from ours that you no longer find Tcl tenable, I for one wish you well with whatever language suits you needs better, and we part friends. -- 73 de ke9tv/2, Kevin
From: Les Cargill on 13 Dec 2009 11:39 Rodericus wrote: > Tk/Tk was my predilect language for small programms, because it was > minimalistic and expresive, low weight and extensible, lisp and C > similar, ideal for embedding it in other programms. Now it is getting > fat and "object oriented" with a lot of unnecesary "features" that > would belong to extensions for special purpose applications. It is > getting a "Cool Programming Language (CPL)" for cool people, not any > more a "Tool Command Language". I think, a splitting and a renaming of > the cool language to something like Cpl/Tk#++ would have been a much > better approach. I think this is the result of having very good > developers not knowing what to do. Please, dont consider this posting > a flame war provocation: it is my oppinion. > > Rodrigo Readi SFAIK, you can still download and use the older releases of the language. -- Les Cargill
From: Arndt Roger Schneider on 13 Dec 2009 12:06
APN schrieb: >To borrow an old slogan from oldsmobile, this indeed is "Not Your >Father's Tcl" (nyf-T, how's that for a new language name - nifty, >no?). But that is a good thing. > >Personally, I feel many/most of the new features greatly expand the >scope and size of software systems you can build using Tcl. So I >suppose that means it is no longer restricted (if it ever was) to be >solely viewed as "tool" language. In that sense, I understand your >point of view but if you feel a language should not "grow" features, >what's preventing you from sticking with older, smaller versions (even >Tcl 7.6) for embedding? > > > [snip] support. |