[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
- To: Joel Voelz -- <joel>
- Subject: Autofact Demo
- From: Mark S. Miller <mark>
- Date: Wed, 16 Aug 89 17:09:46 PDT
- Cc: <marcs>, <xanadu>, <acad!HQ!chrisr>
- In-reply-to: <Joel>,03 PDT <8908162002.AA22838@xxxxxxxxxx>
Date: Wed, 16 Aug 89 13:02:03 PDT
From: joel (Joel Voelz -- "Marketing-in-a-Drum")
It is definitely worth the effort to get Greg "deeply" involved in getting
something working with Xanadu. My additional thoughts are:
1. Are we sure that whatever this demo requires is "on the road" to a full link
between the products?
The amazing answer is yes! The reason I asked the question about "is
it worth slipping Beta by a week to do this?" is that that is my
not-quite-worst case estimate of how much time we'd end up spending on
this that isn't on a direct path to shipping product.
Actually, the amazing part is that the mechanism we'd be using for
connecting between the worlds is the semanically correct ESpace
mechanism. This means that any links made to AutoCad objects now
remain connected in the future when Xanadu becomes the AutoCad
database. In addition, when Xanadu provides spacial coordinate spaces
"box (cat_Float, 3)" and this is used to provide a future AutoCad with
spacial retrieves of parts of a Cad document, then links made with
what we do for AutoFact will still remain connected. I am quite proud
of this, because this transparent preservation of connectivity across
upgrades in capability is where the rubber meets the road on the
semantics. This is the first opportunity I've had to reexamine these
issues with a concrete example, and it all seems like it should work.
The especially pleasing thing is that we'd do the semantically correct
thing for October because no known alternative is easier, even as a
2. Is what we will be able to demonstrate significantly different from the
Depends on what you mean by significantly. The major novelty is
simply the ability to link to & from individual graphical objects
*within* a CAD document, and follow links in both directions. Also,
as far as internal milestones & perceptions are concerned, we'll all
know that it's in many ways "the real thing".
3. WIll we be able to create text documents with the front end and "link" these
documents to AutoCAD drawings? I fwe can do this we might be able to create a
"project management" scenario for the manufacturing of a part - complete with
product specification sheets, manufacturing schedules, memos, etc - that gave
a picture of how the system could really be used. I'm sceptical that just
showing "links" between drawings will be sufficiently "ohh-ahh".
We should be able to do all this, if I understand correctly. Isn't
the only technological requirement for the "project management"
scenario "showing 'links' between drawings" and text, and following in
4. If we can actually implement an elementary inclusion list I say we show it!
Can anybody else actually DO inclusion lists, even if they have seen them? Be-
sides, we have told nearly anyone who listened to us that one of the NEAT fea-
tures of Xanadu is the ability to create Virtual Documents. Well, let's show it
if we can. It would play well in to putting together a "project Management"
demo - where the same drawing or text document could be put in several lists to
be referenced in different contexts.
I believe this is out, as the inclusion list oriented front-end is now
Mac-only, and the Sun front-end is flat-text oriented.
5. Does "basically, getting links to/from AutoCAD" mean only "basic" links or
does it mean link types and filtering???
link types and filtering as far as the back-end & docs and links
technology is concerned. I have no idea what the Sun front-end will
be able to make accesible in this short a time frame. What is left
out is Xanadu style history & version tracking.
Bottom line, I say let's go for it, but let's make sure what we we be able to
show has significant "Ooh-ahhness" so we don't get a "is that all they have done
in a year" comment.
Yeah! Let's go for it! (IMHO)