barnett@crdgw1.crd.ge.com (Bruce Barnett) (12/23/89)
I added comp.windows.misc to the discussion. In article <7391@ficc.uu.net>, peter@ficc (Peter da Silva) writes: >In article <18637@bellcore.bellcore.com> sjs@bellcore.com (Stan Switzer) writes: >> Yes, simplifying models are quite useful, but they are >> not always appropriate to the task at hand. This is why an >> object-oriented interfacing approach is a Good Thing. >But it makes your programs huge. The program itself should have as >little of the user interface as possible as an integral part of itself. I think there is a major difference of perspective here. Some people want environments that let them *build tools* as fast as possible If you can build the basic tool in a few hours, who cares if it's portable? Others want the *tools* to be as fast (and small) as possible. The target market is different, i.e. PC's vs. Workstations. I don't thing the two problems can have have one solution in the near term. Unless you make a PC like a workstation, with all of the extra costs, etc. Another solution Peter is looking for is something more portable than X. >If you need to go to a lower level, do it. But first make the easy things >easy. "Hello World" in a windowing system shouldn't be more than 10 >lines of code. Window systems have evolved so much in the last 10 years. However it is still changing. SunWindows failed the Hello World test. SunView is a better job. Xlib also failed, but the X intrinsics fixes that. (See David Rosenthal's paper in core.src/doc/HelloWorld). Window systems evolve. Tool kits get better. >Oh, and yes. I would LOVE to play around with Display Postscript. But it's >never going to show up in today's PCs, and the way things are going it's >gonna be a long time before it'll show up in any. >Standards aren't just for the elite. The problem the PC's users are facing is evolution. Remember when some "expert" said that 64K was enough for any application? The future PC will be a 486 with 4 Megs of memory for under $1000. You'll see X and NeWS on PC's once some vendor ships Unix V.4. I know there is no standard User Interface or library. But there is a lot more that is not standard. What about the imaging model? Bitmaps are not going to be the model used in the future. Already everyone is talking about outline/rescalable fonts. The PostScript model is even more powerful, and allows for hardware accelerators. Think about the job a bitmap programmer has to do to stretch and rotate a diagram. This is trivial in PostScript. How about the basic model of the user interface? People don't want stand-alone applications. They want to drag a file and drop it onto a tool. Or select several files or objects, and then start the application with the input being the items selected. They also want to cut some information from one tool, and paste in into another tool - like a column of numbers into a chart drawing tool, which in turn is inserted into a document. If you change the data in the table, the chart will be updated and so with the document. These sort of issues make a large difference in the style of programming used to develop the application. What is 10 lines using one model might be 10 pages in another. Selecting a standard now won't solve those problems. Sure, you'll have a product next year, but the year after that you will have to redesign it anyway. I'm sorry for ranting on and on. My old company was very short sighted. They might still be around if they decided to take a few short term risks that had long term benefits. Perhaps that is why I believe so much in new technology. Happy Holidays, everyone. -- Bruce G. Barnett <barnett@crd.ge.com> uunet!crdgw1!barnett
jesup@cbmvax.commodore.com (Randell Jesup) (01/03/90)
In article <4355@crdgw1.crd.ge.com> barnett@crdgw1.crd.ge.com (Bruce Barnett) writes: >The problem the PC's users are facing is evolution. Remember when some >"expert" said that 64K was enough for any application? The future PC >will be a 486 with 4 Megs of memory for under $1000. You'll see X and NeWS >on PC's once some vendor ships Unix V.4. Or maybe an '030 or '040 with 4 Meg... BTW, we (Commodore) already have announced (and demoed in NYC) Sys V.4 for the Amiga 2500 (admittedly not $1000 yet). X is also available for the Amiga and shipping (doesn't require Unix, or even an '020/'030). -- Randell Jesup, Keeper of AmigaDos, Commodore Engineering. {uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.cbm.commodore.com BIX: rjesup Common phrase heard at Amiga Devcon '89: "It's in there!"