Re: The semicolon habit (was: Q: Definition

mitchell@mdd.comm.mot.com (Bill Mitchell) writes:
>Inferring flow control from indent level just does not work in
>practice in a multi-programmer, multi-editor multi-display-format
>situation.


Maybe. But you've listed a lot of constraints that all have to be
true _simultaneously_ for the paradigm to break down. In the scenario
you described, you're going to have maintainability problems, regardless
of how blocks are delimited.


And it's not clear that *scripting* languages will always be used in
the manner you describe. You seem to want to transfer the way you use
a traditional/compiled delivery language like C++ or Ada over to scripting
languages verbatim.


But they have very different roles, and imply very different development
modes. Scripting languages are optimized for speed-of-development, not
complex development scenarios. It's a different kind of tool.


Someone else warned about printing out large programs that follow indentation
rules. But won't you have the same problem for a consistently-indented C
program? The situation is actually better in very-high-level scripting
languages-- programs are smaller, as a rule. If functions start spanning
pages in such languages, it's probably time to rethink your coding practices.


This topic seems destined to become a religious debate. The only thing
I'd suggest is that people curious about this sort of syntax paradigm
have a look at the Python language (see my prior post). People are using
it and liking it; and so far, no reports of problems in multi-programmer
projects, printing code,...


Whether it's Python or not, I'd encourage you to actually look at a language
that uses this model, before falling into some default reaction to new ideas.
At least in the Python world, there's a common experience: it seems unusual
at first glance, but people invariably come to love it after use.


YMMV,
Mark Lutz
--

Return to the comp.compilers page.
Search the comp.compilers archives again.