[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]

Re: [zzdev] "Virtual" (was: Re: [zzdev] Some thoughts on virtualstructures)

Hi Brent,

sorry about not getting back to you earlier!

Brent Turcotte wrote:
> >To get back to the article, it is my understanding that it uses the term
> >in this latter sense: the computer shows "virtual" cells that represent
> >*external* data: a web page as retrieved through HTML; the contents of a
> >database as queried through SQL. So when you used "virtual structures"
> >in conjunction with file formats, I thought you might be talking about,
> >similarly, representing the contents of a file as seen through the file
> >system, not stored in the hypergrid.
> If the cells are "virtual" how can they be edited?  I had assumed it would
> be necessary to store the contents in the hypergrid, otherwise, when an
> edit is made, GZigZag would be forced to modify the original file.  Maybe
> my assumption is wrong.

No, that's correct, and I agree that it would be better to store the
contents in the hypergrid itself. I just wasn't clear whether you meant
this at first, but I think we understand each other now ;o)

> What is better, a physical mapping (storing data in the hypergrid) or a
> virtual
> mapping (temporary storing in virtual cells)?
> I suppose a physical mapping would be in a sense a copy, something that the
> Xanadu project has been trying to avoid (in multiple ways -- the zigzag
> structure itself, transclusions, etc).

Well, yes, but the point was not to lose connections, and that requires
permanent IDs-- which files do not have. So I do not see much in favor
of using a virtual representation of files.

An existing, huge and/or continuously changing store of data can be a
different matter-- which is, AFAIK, what the nutshell article refers to.
For example, if you have an established database service for something,
you would want to explore it inside ZZ, but you do not necessarily want
to download it *as a whole*.

> >I cannot see a good reason to use the term "*virtual* structures" for
> >"mapping of data onto a hypergrid implementation," and I would propose
> >to simply use "hypergrid structure" instead; does this capture your
> >meaning? Or am I missing some good reason to use "virtual structure" instead?
> Hmmm, it seems I used the term 'virtual structures' in too broad a sense for
> too many different things.  How about the term "hypergrid mapping"?

Sounds good.

Any news about the file format -> hypergrid mappings?

- Benja