On Dec 8, 2009, at 10:17 AM, William Stein wrote:
> On Tue, Dec 8, 2009 at 9:06 AM, Robert Bradshaw
> <rober...@math.washington.edu> wrote:
>> On Dec 8, 2009, at 4:49 AM, Stan Schymanski wrote:
>>
>>> Robert Bradshaw wrote:
>>> [SNIP]
>>>> Making notebook IDs that are not simply consecutive
>>>> integers would solve nearly all of your issues above, and I think a
>>>> lot of people (myself included) would appreciate that. Either short
>>>> names or globally unique identifiers (or some combination of both)
>>>> would be a step forward.
>>>>
>>>>
>>> I remember an earlier discussion of this, which I think concluded
>>> that
>>> we want to keep the worksheet names independent of file system
>>> naming
>>> conventions.
>>
>> Yes. Ideally they could be stored in a database as well as stored
>> on a
>> filesystem. It would be nice if it were easy to query, even from
>> outside a running notebook session.
>>
>>> I would also be in favour of unique, static directory names
>>> that show up in the front end. The user could still add more
>>> descriptive
>>> names in addition to the directory names, but it would be nice if we
>>> could refer to other worksheets by their names. The next step would
>>> be
>>> to create scripts that check any cross-references if a worksheet is
>>> re-named; probably not an easy task (?). And if we are at that, I
>>> would
>>> really appreciate a way of referring to particular cells in a
>>> worksheet
>>> by their labels (e.g. \ref{ws:mass_balance1, cell:dMdt}). Maybe a
>>> script
>>> that goes through a whole notebook and checks cross-references in
>>> all
>>> work sheets could then also convert such labels to consecutive
>>> numbers,
>>> similarly to LaTex (?).
>>
>> If worksheet IDs under the hood are globally unique random numbers,
>> and a concordance of name -> ids are kept, one wouldn't have to worry
>> about doing a grand find-and-replace if a name changed.
>
> That's because you would be disallowing name changes.
> Making users refer to a random id number whenever they want to refer
> to or reference a given worksheet is kind of mean and also makes
> autogeneration of collections of worksheets by other programs more
> difficult. It would be like forcing latex users to *remember* funny
> random id's instead of using \ref{sec:intro}, which is much easier.
>
> There will be unique id's, just as there are now, but that should be
> something the user doesn't have to worry about.
I think you misunderstood my proposal--all the user would see is the
friendly name, which the user could change. The ID would just be
stored internally. It would help with situation where multiple
worksheets might have the same name (e.g. say I sent you a bundle of
worksheets "heegner" and "kolyvagen," where the latter referenced the
former, but you already had a worksheet with name/id "heegner." It
could also handle if, the next day, I sent you an updated "kolyvagen"
and when you uploaded it it would link to the correct "heegner" (or
whatever you had to rename it to)). There's also the situation where I
could be working on a set of worksheets on my own machine and then
uploading them to a public server later to share. If I changed a label
on my own machine, it could do a find-and-replace there, but would of
course break cross references on the server unless I re-uploaded the
whole connected component.
I guess what I'm advocating is that it would be useful for the
underlying ID to be (probabilistically) globally unique, not just
unique for a user or a specific notebook server.
- Robert
--
To post to this group, send email to sage-support@googlegroups.com
To unsubscribe from this group, send email to
sage-support+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org
