4 ms·

TCL is dead)))

4 points by alexkarta 11y ago

dalke 11y ago
Will you explain what I'm looking at? How does that page show that 'TCL is dead'? As far as I can tell, it means that Tcl developers do not use Debian to install debug builds and documentation.
- alexkarta 11y ago
  I thought, it is quite clear from the graph that tcl, despite of many pessimists, is alive. It is still popular and has positive tendencies. Now, tcl 8.6 is more popular, then tcl 8.5 was (https://qa.debian.org/popcon.php?package=tcl8.5 https://qa.debian.org/popcon.php?package=tcl8.5). I understand that popcon is not perfect for such estimations, but I feel the same tendencies.
  - dalke 11y ago
    If I understand it correctly, you used a counter-factual title 'TCL is dead' (note also the non-convention capitalization - it should be 'Tcl'), and assumed that a good number of people on HN would understand 1) it was satirical, 2) the need for counter-pessimists, 3) the context in which to understand the numbers shown, and 4) whatever 'popcon' means. I - and I'll add that I've published papers on using Tcl for molecular graphics, in software which still has 10K+ users - don't have any of that understanding. I know Tcl is still active, if a minority language, and is a good fit for some niches. (One of which is regression testing, with the SQLite test suite being perhaps the most famous.)
    - alexkarta 11y ago
      Ok, thank you. "10K+ users" is great and promising fact. After 5+ years of active python developing, I have decided to move to Tcl (you are right, "TCL" looks really wrong). I feel this language is fit to my tasks much better then python
      - dalke 11y ago
        I haven't worked on the project for a while. I pulled up the 2011 survey report at http://www.ks.uiuc.edu/Research/vmd/survey/report2011/vmdsurvey2011rep.pdf http://www.ks.uiuc.edu/Research/vmd/survey/report2011/vmdsur... . There were 56,575 registered users then. Mind you, I don't know how many of those people write scripts. Plus, VMD also supports Python, though that was added years after Tcl. Personally, I think Tcl is better for the types of scripting that the physical scientists (biologist, chemists, etc.) - most of my user base - want to do. However, as something they developed became more generally useful, I found I couldn't easily tweak it to be part of the larger system. OTOH, while Python is more complex than Tcl, I rarely feel the need to rewrite user code in Python into C++ code, so there's a more smooth transition between non-programmer code to experienced programmer code. That is one of the reasons I switched to Python. On the flip side, IPython and others have developed script-like meta-languages which makes it easier to do some of the things that Tcl excels at.
    - SQLite 11y ago
      Indeed, SQLite is merely a Tcl extension that "escaped" into the wild. SQLite no longer depends on Tcl to run, but it does still depends on Tcl for its make process and for testing. The SQLite development team also depends on tools written in Tcl/Tk for tasks such as editing and collaboration. Where Tcl to die, SQLite would die with it. Fortunately, both outcomes seem unlikely. Some of us are old enough to remember the frequent prognostications that "unix is dead" and was being replaced by more modern OSes like VMS or OS/2. How'd that work out?

4 points by alexkarta 11y ago

dalke 11y ago

Will you explain what I'm looking at? How does that page show that 'TCL is dead'? As far as I can tell, it means that Tcl developers do not use Debian to install debug builds and documentation.

alexkarta 11y ago
I thought, it is quite clear from the graph that tcl, despite of many pessimists, is alive. It is still popular and has positive tendencies. Now, tcl 8.6 is more popular, then tcl 8.5 was (https://qa.debian.org/popcon.php?package=tcl8.5 https://qa.debian.org/popcon.php?package=tcl8.5). I understand that popcon is not perfect for such estimations, but I feel the same tendencies.
- dalke 11y ago
  If I understand it correctly, you used a counter-factual title 'TCL is dead' (note also the non-convention capitalization - it should be 'Tcl'), and assumed that a good number of people on HN would understand 1) it was satirical, 2) the need for counter-pessimists, 3) the context in which to understand the numbers shown, and 4) whatever 'popcon' means. I - and I'll add that I've published papers on using Tcl for molecular graphics, in software which still has 10K+ users - don't have any of that understanding. I know Tcl is still active, if a minority language, and is a good fit for some niches. (One of which is regression testing, with the SQLite test suite being perhaps the most famous.)
  - alexkarta 11y ago
    Ok, thank you. "10K+ users" is great and promising fact. After 5+ years of active python developing, I have decided to move to Tcl (you are right, "TCL" looks really wrong). I feel this language is fit to my tasks much better then python
    - dalke 11y ago
      I haven't worked on the project for a while. I pulled up the 2011 survey report at http://www.ks.uiuc.edu/Research/vmd/survey/report2011/vmdsurvey2011rep.pdf http://www.ks.uiuc.edu/Research/vmd/survey/report2011/vmdsur... . There were 56,575 registered users then. Mind you, I don't know how many of those people write scripts. Plus, VMD also supports Python, though that was added years after Tcl. Personally, I think Tcl is better for the types of scripting that the physical scientists (biologist, chemists, etc.) - most of my user base - want to do. However, as something they developed became more generally useful, I found I couldn't easily tweak it to be part of the larger system. OTOH, while Python is more complex than Tcl, I rarely feel the need to rewrite user code in Python into C++ code, so there's a more smooth transition between non-programmer code to experienced programmer code. That is one of the reasons I switched to Python. On the flip side, IPython and others have developed script-like meta-languages which makes it easier to do some of the things that Tcl excels at.
  - SQLite 11y ago
    Indeed, SQLite is merely a Tcl extension that "escaped" into the wild. SQLite no longer depends on Tcl to run, but it does still depends on Tcl for its make process and for testing. The SQLite development team also depends on tools written in Tcl/Tk for tasks such as editing and collaboration. Where Tcl to die, SQLite would die with it. Fortunately, both outcomes seem unlikely. Some of us are old enough to remember the frequent prognostications that "unix is dead" and was being replaced by more modern OSes like VMS or OS/2. How'd that work out?