home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.advocacy
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!eff!news.byu.edu!ux1!fcom.cc.utah.edu!jaguar.cs.utah.edu!brian
- From: brian@jaguar.cs.utah.edu (Brian Sturgill)
- Subject: Re: Some new tidbits and my 'retirement' from os2.advocacy.
- Message-ID: <1992Nov6.235129.24616@fcom.cc.utah.edu>
- Sender: news@fcom.cc.utah.edu
- Organization: University of Utah CS Dept
- References: <BxBBnv.BAn@news.iastate.edu>
- Distribution: usa
- Date: Fri, 6 Nov 92 23:51:29 GMT
- Lines: 86
-
- In article <BxBBnv.BAn@news.iastate.edu> TW.FY4@isumvs.iastate.edu (Timothy I Miller) writes:
- ...
- >In article <1992Nov6.080429.10588@fcom.cc.utah.edu>,
- >>[PC Week; November 2, 1992; Page 57,64]
- >>[A reminder from my OS/2 developer conference report:
- >> Mirrors leaves the app code itself as 16-bit.
- >> Apps that use mirrors incur that overhead, the overhead
- >> of the API translation, and the overhead of the thunking
- >> from the 16-bit to 32-bit DLLs.
- >> Plus they are really written for Windows, not OS/2.
- >>]
- >
- > [Stuff about Mirrors deleted...]
- >
- >>...
- >>[
- >>Well, this is good news for the Describe folks, neither WordPerfect
- >>nor AmiPro will be true 32-bit apps.
- >>]
- >
- > Well, as I understand it (and I heard it from Lotus), Ami Pro
- >and 123 are both going to be completely native 32-bit apps. This is
- >because they use Borland C/C++ to do the development. I don't know
- >about the rest of their products, though.
-
- Well according to the Micrografx representative at the OS/2 developers
- conference, the following products/vendors will be using Mirrors:
- Micrographix (Draw and Designer -- Draw was released the day before
- the conference.)
- Corel (Draw)
- Spinnaker (PFS: Works)
- Zsoft (Publisher's Paintbrush)
- Lotus (Most apps, including Ami Pro)
- Borland
- Interleaf
- Assymetrix
-
- Micrografx represents the product that you come out with as a 32-bit app
- but the truth is that the code of the app itself is 16-bit. When the
- app makes a Windows call, it thunks to a 32-bit DLL that emulates the
- Windows call under PM.
- Thus Lotus may call Ami Pro a 32-bit OS/2 2.0 app (and it will only run
- under OS/2 2.0), but performance-wise, and especially feature-wise it is far
- from one.
-
- The method used will cause an app like a word processor to possibly run
- faster than under Windows 3.1, but definitely not significantly so.
-
- An app like a spreadsheet, where the 16-bit part of the app does non-trivial
- CPU-oriented things, could easily end up running slower than under
- Windows 3.1.
-
- While Micrografx said that their new release of Draw for OS/2 was slightly
- faster than the Windows 3.1 version, the evidence I saw said otherwise.
- The guy was using a 486 33 (I asked) and it was definitely responding
- slower than Windows Draw does for me under 3.1 on a _386_ 33. This of course
- is not a very good test as we had different video hardware. If anyone out
- there is in a position to make an objective test, I'd be interested in the
- results.
-
- In any event the apps made this way will run slower than apps even
- simply recompiled for Win32.
-
- Once Micrografx gets the Win32s version done, then these apps will run
- as 32-bit throughout, and thus only incur API translation overhead.
-
- However there are three things abundantly clear:
- Most if not all the big names are sticking with the Windows API.
-
- Once Micrografx gets the Win32s version out, why would mainstream
- app vendors want to make PM apps?
-
- OS/2 apps written to this way will always lag the Windows versions.
-
- Thus is it any wonder that IBM broke off that part of its relationship with
- Micrografx?
-
- >
- >Timothy Miller
-
- Brian
- --
- C. Brian Sturgill I strongly encourage you to mail your copy
- University of Utah of Ross Perot's: "United We Stand
- Center for Software Science How we can take back our country",
- brian@cs.utah.edu; CIS: 70363,1373 to President-elect Clinton.
-