home *** CD-ROM | disk | FTP | other *** search
- Subject: Fwd> Re> OSA Cross-platform Issues
- Sent: 7/10/96 4:25 PM
- Received: 7/10/96 4:41 PM
- From: David McCusker, david_mccusker@powertalk.apple.com
- Reply-To: ODF Interest, ODF-Interest@CILabs.ORG
- To: OpenDoc Development Framework Discussion List, ODF-Interest@CILabs.
-
- Christopher Hunt (chris_hunt@yes.optus.com.au) writes:
- > I've been curious about OpenDoc's strategy for inter-part/task
- > communication on non-Mac platforms... could someone please elaborate
- > more on Gregs' points i.e. how will an OpenDoc part editor be expected
- > to communicate with other parts and tasks; simply through proprietary
- > interfaces?
-
- Jon Pugh had something pretty interesting to say about cross platform
- OSA on the OpenDoc interest list earlier this afternoon. I'm forwarding
- it to this list (it follows my signature).
-
- Additionally, one might also note that OpenDoc's inter-part communication
- strategy allows for parts to roll their own without interference from
- OpenDoc. We recommend using extensions to advertise your brand of
- inter-part communication. Examining extensions requires making SOM calls.
- Waiting for DSOM simplifies the inter-process case for communiating with
- SOM objects.
-
- David McCusker, OpenDoc Storage & Bento, david_mccusker@powertalk.apple.com
-
- <JON PUGH>
- At 11:55 AM 7/10/96, Greg Friedman wrote:
- >A cross platform scripting implementation in ODF is conditional on
- >a cross platform scripting implementation in OpenDoc.
-
- Greg is completely correct here, but then, that wasn't the issue. ;)
-
- It was initially hoped that we could get OSA ported as part of OpenDoc, but
- that hasn't happened. It did get ported to OS/2 and the OSA is present
- there using REXX as their scripting language, but it has not made it to the
- Windows port and that doesn't look likely to happen, so we're pretty much
- screwed as far as the OSA is concerned.
-
- What has happened is that OLE Automation is the dominant scripting
- architecture on Windows (surprise surprise surprise). What we need then is
- a way to reconcile OSA (as used by AppleScript and other Macintosh
- scripting languages) and other alternative messaging systems, such as OLE
- Automation, Denali and DSOM.
-
- What I've been doing is putting together a plan for replacing the messaging
- subsystem with an object based system which would reconcile these disparate
- views of the world.
-
- Basically, we are creating a new SDL (Scripting Definition Language) which
- describes your part's scripting interface in terms of objects. This is an
- absurdly simple language that includes some specific constructs for the
- various languages we need to support. We are writing a parser which will
- read this and dump .idl, .cpp and .r files for a variety of
- implementations. The first one we're doing is OSA with Denali right behind
- it and OLE coming in later. The beauty is that all of these
- implementations will bottleneck on a single .cpp file per scripting object
- which developers will fill in with their implementation. This will be a
- simple SOM object. All the other files merely call this base
- implementation and translate between their general implementations and this
- specific one. In addition, the parser will dump a complete aete resource
- for your part.
-
- All of this is a work in progress with no timelines specified yet, but it's
- looking very good in design and there's actual work being done on it. It's
- going to include a compatibility object which knows how to talk to parts
- using the current semantic interface, so there's no danger in losing
- functionality by implementing things now, but it will be a good deal easier
- to cope with all of theses issues when this is finished.
-
- I don't know any of the issues regarding getting this accepted by IBM for
- the other versions, but technically, they were asking for this and it could
- probably be done fairly easily since it looks like the OSA is the most
- work.
-
- At 2:03 PM 7/10/96, Bernie Wieser wrote:
- >Consider implementing an OpenDoc extension that exposes JUST THE CONTENT of
- >your part via SOM. We did. It's not recordable, but it is fast
- >("direct-to-som").
-
- This is essentially what we are doing, plus we are automagically making the
- glue for all the various messaging systems at the same time.
-
- I just wish I could snap my fingers and have it all working. ;)
-
- Jon
- </JON PUGH>
-
-
- ------------------ Nested Letter Follows ------------------
- At 11:55 AM 7/10/96, Greg Friedman wrote:
- >A cross platform scripting implementation in ODF is conditional on
- >a cross platform scripting implementation in OpenDoc.
-
- Greg is completely correct here, but then, that wasn't the issue. ;)
-
- It was initially hoped that we could get OSA ported as part of OpenDoc, but
- that hasn't happened. It did get ported to OS/2 and the OSA is present
- there using REXX as their scripting language, but it has not made it to the
- Windows port and that doesn't look likely to happen, so we're pretty much
- screwed as far as the OSA is concerned.
-
- What has happened is that OLE Automation is the dominant scripting
- architecture on Windows (surprise surprise surprise). What we need then is
- a way to reconcile OSA (as used by AppleScript and other Macintosh
- scripting languages) and other alternative messaging systems, such as OLE
- Automation, Denali and DSOM.
-
- What I've been doing is putting together a plan for replacing the messaging
- subsystem with an object based system which would reconcile these disparate
- views of the world.
-
- Basically, we are creating a new SDL (Scripting Definition Language) which
- describes your part's scripting interface in terms of objects. This is an
- absurdly simple language that includes some specific constructs for the
- various languages we need to support. We are writing a parser which will
- read this and dump .idl, .cpp and .r files for a variety of
- implementations. The first one we're doing is OSA with Denali right behind
- it and OLE coming in later. The beauty is that all of these
- implementations will bottleneck on a single .cpp file per scripting object
- which developers will fill in with their implementation. This will be a
- simple SOM object. All the other files merely call this base
- implementation and translate between their general implementations and this
- specific one. In addition, the parser will dump a complete aete resource
- for your part.
-
- All of this is a work in progress with no timelines specified yet, but it's
- looking very good in design and there's actual work being done on it. It's
- going to include a compatibility object which knows how to talk to parts
- using the current semantic interface, so there's no danger in losing
- functionality by implementing things now, but it will be a good deal easier
- to cope with all of theses issues when this is finished.
-
- I don't know any of the issues regarding getting this accepted by IBM for
- the other versions, but technically, they were asking for this and it could
- probably be done fairly easily since it looks like the OSA is the most
- work.
-
- At 2:03 PM 7/10/96, Bernie Wieser wrote:
- >Consider implementing an OpenDoc extension that exposes JUST THE CONTENT of
- >your part via SOM. We did. It's not recordable, but it is fast
- >("direct-to-som").
-
- This is essentially what we are doing, plus we are automagically making the
- glue for all the various messaging systems at the same time.
-
- I just wish I could snap my fingers and have it all working. ;)
-
- Jon
-
-
-
- ------------------ RFC822 Header Follows ------------------
- Received: by powertalk.apple.com with SMTP;10 Jul 1996 13:46:57 -0700
- Received: from CILabs.ORG ([199.182.213.62]) by mail-in1.apple.com
- (8.6.12/8.6.12) with SMTP id NAA17985 for
- <david_mccusker@powertalk.apple.com>; Wed, 10 Jul 1996 13:37:07 -0700
- Received: from localhost by CILabs.ORG (5.x/SMI-SVR4) id AA16352; Wed, 10 Jul
- 1996 13:42:12 -0700
- Received: from mail-out1.apple.com by CILabs.ORG (5.x/SMI-SVR4) id AA16315;
- Wed, 10 Jul 1996 13:35:35 -0700
- Received: from [17.203.32.142] (A17-203-32-142.apple.com [17.203.32.142]) by
- mail-out1.apple.com (8.7.5/8.7.3) with SMTP id NAA03814 for
- <OpenDoc-Interest@CILabs.ORG>; Wed, 10 Jul 1996 13:34:09 -0700
- Message-Id: <v02140b08ae09b99b18c7@[17.203.32.142]>
- Date: Wed, 10 Jul 1996 13:33:50 -0700
- Reply-To: OpenDoc-Interest@CILabs.ORG
- Sender: owner-OpenDoc-Interest@CILabs.ORG
- Precedence: list
- From: jonpugh@netcom.com (Jon Pugh)
- To: OpenDoc Related Technologies Interest List <OpenDoc-Interest@CILabs.ORG>
- Subject: Re: OSA Cross-platform Issues
- Mime-Version: 1.0
- Content-Type: text/plain; charset="us-ascii"
- X-Sender: jonpugh@netcom3.netcom.com
- X-Url: http://www.infoworkshop.com/~jonpugh/
- X-Listprocessor-Version: 7.1 -- ListProcessor by CREN
-
-
- ------------------ End of Nested Letter ------------------
-
-