On Tue, Jan 3, 2012 at 3:48 PM, Seth Johnson <[email protected]> wrote:
> On Tue, Jan 3, 2012 at 3:20 PM, Ville M. Vainio <[email protected]> wrote:
>> On Tue, Jan 3, 2012 at 6:41 PM, Terry Brown <[email protected]> wrote:
>>
>>> I think that's the case.  @<file> handling is complex, I'm not sure how
>>> familiar with it Ville is, I've poked around the edges a bit, but don't
>>> know how it would interact with loading to / from a DB.
>>
>> My code just dumps everything, including under @<file> stuff, to db.
>>
>> This is by design, as I want to ship .leoq as a standalone file that
>> you can send to your phone in email, or whatever.
>>
>> Not writing @file nodes is easy, just stop traversal when you see one.
>> Likewise, reading back in is easy, just expand the whole tree and
>> after that do the @file node handling for the whole tree.
>>
>>> And even if that wasn't an issue, the next step beyond using a DB to
>>> replace the file system would also be complex, whether it was sharing
>>> or whatever.  Well, perhaps versioning wouldn't be so hard, but
>>> sharing is certainly challenging.
>>
>> I think it's best to left advanced use cases like that unhandled. Even
>> if coding it wasn't too hard (it probably is ;-), people want simple
>> and reliable workflows for their creative work, users are paranoid
>> about losing data if they can't completely understand what is
>> happening under the hood (e.g. clone wars must have spooked of many of
>> us)
>
>
> I think we should leave everything new unhandled.  Just regard the db
> as a file system you're saving your own Leo files into.  The only new
> thing should be, instead of just saving it as one file as a whole,
> store the elements as separate units (records, nodes, whatever), but
> just make sure that you have a save and load procedure for that db
> representation, that works because all it does is save your file in
> that form and load it.  Then everything current in Leo will just work.
>  Plus people would know what they need to store in the db
> representation, and what they need to feed to the load procedure.
> They can use that and hack at different ideas, just knowing that at
> bottom the classic, basic Leo is still going to work so long as they
> represent it equivalently to that reference representation, and feed
> the load procedure with what it needs.

And why not, for the @file business, keep those saving to the local
standard file system: just put the .leo file in the db for now.  In
the routine for saving, have a flag that once it's set, will save the
.leo file in the db, as well as save it in the traditional fiel system
-- or one or the other.  I don't see why we have to start by solving
distributed, collaborative work on the @file stuff, with versioning.
Just save the .leo file.  Let people suggest different approaches
starting from there.

(Getting redundant, sorry.  I'm done.  :-)  )


Seth

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/leo-editor?hl=en.

Reply via email to