Chris Burke wrote:
I am in two minds as to whether anything is required here. However, I do
think that delete all temporary files would be the more useful option, as
rarely do I want to delete a specific temporary file. After a bit, the
temp directory is filled with junk, and you just want to clean the whole
thing up.
Anyway, I put a File|Delete Temporary Files menu item in the latest
scripts, and we can see if it turns out to be useful.
For me, the temp file is a convenient immediately accessible file which has
lots of small, often one off tasks or snippets which I may want to refer
back to. I normally write a first line which provides me enough detail
about the task for later reference. If I begin to use material in a script
more frequently I save it with a name but in the temp directory. As an aide
memoir I have a utility, firstline'' which displays the file name and
firstline for all scripts in temp as default, but works for any directory.
I think the J environment has been developed with a primary focus on
programming and system development, not on its use as an interactive working
environment. The project manager is great for managing a system development
project, but is not oriented towards a data analysis project, which is the
sort of thing I am commonly engaged in.
For that, having the simple accessible system provided by storing and
accessing stuff in \temp is invaluable. Some stuff stays there quite long
term. Other bits get embedded in a system, or at the end of a project get
filed away as a record of what was done. Having the ability to flexibly
modify the current folder used for such a \temp storage location would be a
useful enhancement. In the dbeta Edit\Configure\Folders enables setting
~user, but that is not the default directory for \temp.
Temp provides a simple default filing system within which there is a
sequential record of the scripts created and used. As such it is extremely
useful. In such an environment, a single file deletion is fine, but erasing
the whole record of such activities could be very costly.
Essentially I want you to think seriously about the needs of users who are
not primarily programmers, who have complex data analysis tasks where the
tree of steps they take evolves as they expand their understanding of the
problem, and where a record of the sequence of their steps is important to
them. For them the present Temp file mechanism is a flexible and valuable
tool. Erasing everything in it is the last thing one is interested in. If
we really want to expand the user base of J we need to address the needs of
such users and not create simple routes to dangerous actions.
Fraser
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
