home *** CD-ROM | disk | FTP | other *** search
- Path: yama.mcc.ac.uk!dmu!mkcsug06
- From: hcm94rp2@dmu.ac.uk (Richard Perrott)
- Newsgroups: comp.sys.amiga.applications
- Subject: Re: MUI
- Date: Tue, 05 Mar 96 23:13:58 GMT
- Organization: De Montfort University
- Message-ID: <4hi0ej$906@macondo.dmu.ac.uk>
- References: <4h22lr$1d8@ns.hookon.be> <singh-0404952044330001@pool5-028.wwa.com> <4hdstb$2aj@ringer.cs.utsa.edu>
- NNTP-Posting-Host: mkcsug06.mk.dmu.ac.uk
- X-Newsreader: News Xpress Version 1.0 Beta #4
-
- Glossary:
- MFU Massive Fuck Up, a better name than 'BUG' for programming
- defects since it forces the programmer to accept
- responsibility for them not just call them a 'feature'.
- The term is taken from a excellent book on bug
- prevention and debugging.
-
- In article <4hdstb$2aj@ringer.cs.utsa.edu>,
- jpeacock@ringer.cs.utsa.edu (Jason Peacock) wrote:
-
- .cut..
- > I really don't get this. I don't see how my 40Mhz 68030 could be
- > faster than your 68040 at any speed. MUI 2.x was a bit slow but
- > still tolerable and MUI 3.x feels about 2-3 times faster, which is
- > nice. The other stuff out there (gtlibrary, BGUI, Triton, and internal
- > mechanisms) doesn't feel any faster than MUI anymore.
- .cut..
- I haven't noticed much if any speed improvement and MUI is the slowest of
- numerous GUI libraries I have tried so far, even MUI V3.3.
-
- >
- >: It really boggles my mind why with a nice, small, fast OS like the Amiga
- >: so many programmers want to tether it to an interface as sluggish an MUI.
- >
- > I've always wondered if people thought MUI was so slow on a nice, small,
- > fast OS like the Amiga, imagine what it would be like if someone tried
- > to implement something similar for Windows or Macintosh.
- >
- They'd accept it since they are used to sluggish GUI's which crash due to
- MFUs, e.g. it needs a Pentium to get acceptable GUI speed for most
- applications in Windows 3.1 and they can still get btrieve crashes.
- By acceptable I mean when they attempt to multitask.
-
- >: It's a cruel twist of fate that MUI has become so popular -- I
- >: dunno, maybe programmers are too lazy to write their own boopsi gadgets?
- >
- > Why re-invent the wheel? If someone's already done it, use it. It
- > does no good to have a "not invented here" attitude. Why waste
- > time building another boopsi gadget if one that meets your needs
- > already exists for use?
- >
- Granted re-writing boopsi gadgets is questionable, but having a _very_ slow
- handler code, MUI, is down right stupid since the client program will be
- annoyingly slow.
-
- >: It beats me, but all in all, after using MUI on my brother's 2000/020 it
- >: makes me feel sorry -- I always thought the amiga community was the one
- >: group that took care not to make older machines obsolete. The "hey if
- >: they user feels the interface is kludgy, he or she needs more horsepower"
- >: theory is way too MicroSoft/Intel for my liking. :(
- >
- > I still think we do. Amiga Technologies is coming out with the
- > Amiga Surfer package which will be a 14Mhz 68ec020 A1200 running
- > Internet software that requires MUI (included) to run. AT must
- > think that the A1200 is up to handling MUI.
- >
- They made a mistake in under-estimating resources before e.g. not enough
- memory for a certain graphics tool on HD. Also they probably wanted to rush
- out some response just to say they were first.
-
- > Besides, there's only so much one can do with limited resources. I
- > remember a program called Paperclip Publisher for the C64. It
- > amazed me with the amount of features they were able to put into a
- > computer that had a single 140KB floppy drive, 64KB RAM, and a 1Mhz
- > processor. I was not amazed, however, that the program ran with all
- > the speed of molasses in January. GEOS for the C64/C128 is another
- > example. These computers were being pushed to their limits by these
- > programs and the speed was barely tolerable.
- >
- > If you want more features, and have them done well, you can't expect
- > the hardware to remain static and still cope with growing software.
- >
- I have an A1200 with 4MB _extra_ Fast Mem, 50MHz 030/MMU, and 1GB HD, not a
- cheap set up. After my KS is cached, buffers for 3 AFS-Pro partitions, and a
- few commodities I only have 1.5MB Chip and 2MB Fast left. I think it is
- reasonable to expect to run most programs in 3.5MB of RAM.
-
- Most programs run comfortably in this environment, but MUI applications are
- still noticably sluggish. For a directory tool e.g. RO where dos is handling
- most of the actual work, for the GUI to be sluggish is a disappointment. RO
- is hardly used now since the fancy front end takes up too much space and some
- MUI objects actually make the interface harder to use than the equivalent
- Gadstools/boopsi one. I have used computers for over 10 years and done a GUI
- design course so I think know what I'm talking about.
-
- As for memory:
- MUI seems to keep RAM for items even when those items are not being used, and
- doesn't flush them until you close and flush MUI. If MUI would flush them
- when it was still loaded then I wouldn't _have_ to do this manually.
-
- Also since MUI and its classes have so many data stored in codes hunks I
- can't use VMM for a lot of the data parts, since MUI code and applications
- won't tolerate being in virtual memory. Most normal applications will take
- advantage of VMM since a lot of their data is in data hunks.
-
- Also despite MUI has all this junk cached in memory it runs slower than
- applications which setup stuctures only _when_required_.
-
- IMHO
- Richard
-
- ----------------------------------------------------------
- Richard Perrott | Amiga Forever!
- hcm94rp2@dmu.ac.uk | If the answer is Microsoft or
- *nobody@REPLAY.COM is in| Apple you're asking the wrong
- my kill file. | question.
-