Case sensitivity, line labels and the future

--- Jay Belanger <belanger@truman.edu> wrote:

> > Basically if we want to make use of case sensitive code but our
> > standard is to have all contributed code be lower case, we have to
> > rewrite it to use only lower case.
> 
> Wouldn't a filter that takes the files and makes them all lower case
> work? 

Not if the proper functioning of the file depends on case sensitivity.
 
> > For example, if someone writes a file that defines a function Diff
> > which does something different from diff, it's a problem when we
> > convert to lower case.  
> 
> But that's creating new files, not rewriting old ones.  And why would
> Diff be converted to lower case?

Right - we're thinking ahead to what users might write and want to
share.  Our files we can control, but what other people write is an
unknown.  Diff would be lower case if we were doing a basic convert of
a file to be lower case.

> > The way things are shaping up, I think we will be either a)
> > autoconvering source files to lower case code and fixing what
> breaks or
> 
> Since the current situation is case insensitive, what could break?
> There can't be any situation where Diff and diff act differently.
> At any rate, this would go to the trouble of making Maxima case
> sensitive.  But given a case sensitive and case insensitive Maxima,
> I've seen several reasons to choose the former.  Is there any reason
> to choose the latter? 

AFAIK, the current situation is not strictly case insensitive, but is
governed by some rather strange rules.  Whatever else happens, we want
to come out of this with a uniform, consistent policy.

No, I know of no good reason for case insensitivity except some people
prefer working in an environment where they don't have to worry about
case.

> > > > Martin RUBEY <rubey@labri.fr> writes:
> ...
> > As soon as we have case-sensitive packages, it will be impossible
> to use
> > them in case-INsensitive maxima. 
> 
> So the concern here is backwards compatibility?

In essence.  That's what my proposal was addressing.  But if we go
strictly case sensitive in Maxima it won't be an issue.

> > So as long as maxima is possibly case-insensitive (as it is now or
> > as it might be in the future by choice), such a package might be
> > broken. (because it might contain two different variables Foo and
> > foo, which would be the same in case-insensitive maxima)
> 
> If Maxima were made case sensitive, why would it revert to case
> insensitivity?  The only reason that I've seen to remain case
> insensitive is the trouble of making it case sensitive in the first
> place.

No, I think a majority agree case sensitivity is a good way to go. My
thought was to transparently allow both, but if that doesn't happen my
vote is definitely for case sensitive. 

> (To be clear, I'm not being purposefully dense.  I'm that way
> naturally.) 

You're not being dense :-).  It's just a tricky issue.  

CY

__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com