home *** CD-ROM | disk | FTP | other *** search
Text File | 1991-03-06 | 3.7 KB | 81 lines | [TEXT/GEOL] |
- Item 4443094 20-Jan-91 09:51PST
-
- From: CHADWICK1 Attain, Dan Chadwick,PRT
-
- To: MACAPP.TECH$ MacApp Technical
-
- Item forwarded by FRED.FORSMAN to MATTSON1
-
- Item forwarded by SPA.DTS to SPA0144
-
- ------------------------------------------------------------------------------
-
- Sub: Ramblings
-
- Greetings-
-
- Having just downloaded package 63 of messages from MacApp.Tech$, I am a bit
- tardy in replying. Please excuse my rambling…
-
- Mike Latta-
- A decade ago Wang had an internal performance analysis tool (called the
- Software Oscilloscope) which "charged back" to the caller the time that a
- subroutine used. I understand it was a wonderful, but limited tool because it
- resulted in several orders of magnitude performance degradation - like 24 hours
- for a small test. I suspect that it charged every instruction, rather than
- sampling.
-
- I too have found the performance reports difficult to analyze. Putting in
- a full crawl, however, might render them too slow to use. I would love to see
- this tried, however.
-
- New MacAppers-
-
- Regarding Think Pascal problems mentioned by Dave Shillito: 1) I too don't
- like having to buy into Think's reformatting. However, after a while, I let my
- zenophobic defenses down and happily traded my indenting style for decreased
- editing time. 2) I have never had Think Pascal crash when working on a virgin
- project file - one that has never been used to run within the environment. I
- keep such a virgin project around and copy it whenever my code has caused Think
- get flakey. I suspect that only running in protected mode, and perhaps not
- even that, can eliminate this problem. Also, Thinks crashes are really no
- worse than having your application crash the whole machine when running under
- MultiFinder with MPW. 3) The modified MacApp source is a pain, but if you don't
- want to be on the bleeding edge of MacApp releases, it is really not a problem
- to wait for Think to release a new converter. I wish that the MacApp/MPW/Think
- teams would make Think source code compatibility a non-issue.
-
- I wrote FileMaker versions 1.0 through roman II with MPW and TMON (and not
- MacApp - it didn't exist in 1984!). I can now find bugs MUCH more quickly with
- Think than I can with TMON, The Debugger, or SADE. Now that I have some MacApp
- experience, I realize that many problems posted on MacApp.Tech$ could easily be
- found with the Think Debugger. I suspect that the people who post these
- messages are not lazy but rather that they are not using the Think Debugger
- (for whatever reason). I personally wish that Apple would buy the Think
- division from Symantec and make it the standard technology for developing
- applications on the Macintosh. I understand that Think is not a big MoneyMaker
- for Symantec and so I worry about its long-term future.
-
- Ed Dry-
-
- I too would hate to see AppleLink cluttered with political messages. I do
- feel that ONE posting of a 900 number for the congressional and white house
- opinion line - a line that may be used for pro and anti sentiment - on a issue
- as important as the Persian Gulf war is appropriate. There ARE more important
- things in life than Macintosh, after all.
-
- James Plamondon-
-
- Regarding your contest, there is a non-portable way to directly call a
- method more than on "inherited level" up in the chain. It is something like
- TControl$Draw. Since I don't use MPW anymore, I can't recall exactly. I had
- this exact some problem, solved it with the above hack, ported to TP3, and then
- copied in the adornment code. Sigh. Were you unaware of this, did you want to
- point out the need for this cabability to the '9X people, or do you have
- another really cool trick up your sleeve?
-
- Dan Chadwick
-
-
-
-