home *** CD-ROM | disk | FTP | other *** search
- Path: news.luth.se!usenet
- From: Fredrik Hellgren <fleck@ludd.luth.se>
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: What the new Amiga-OS *must* have
- Date: Fri, 22 Mar 1996 22:42:06 +0100
- Organization: University of Lulea, Sweden
- Message-ID: <31531EAE.2781E494@ludd.luth.se>
- References: <4iqgd2$8n7@newsbf02.news.aol.com> <DoM13v.CrB@iglou.com> <4it1c3$fts@ar.ar.com.au>
- NNTP-Posting-Host: my21.sm.luth.se
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.0 (X11; I; SunOS 4.1.4 sun4)
-
- Storm/Cydonia wrote:
- >
- > Steve L. Edford (sledford@iglou.iglou.com) wrote:
- > : cmarschn@aol.com (Cmarschn) writes:
- >
- > : >- A fat binary format is badly needed. Not only that it could hold both
- > : >680x0 and PowerPC-Code, the most important thing that is
- >
- > : This sounds very reasonable....matter of fact, I like it! :-) Well, at
- > : least the part about bridging between 68K & PowerPC code. Unfortunately,
- > : this will probably mean more bloated code.
- >
- > This I do not understand. Why on earth is it better to have a fat binary,
- > containing both sets of code, and therefore making the executable twice
- > as big, with half of it NEVER being refereced. (OK not twice as big, 'cos
- > the data would only be there once, but bigger). Why is this better than
- > simply having two executables, one for 680x0 and one for PPC, and letting
- > the user install the one he needs, without the other one's unnecessary
- > code cluttering his hard disk.
- >
- > Maybe someone with more knowledge of these fat binaries could tell me
- > what the advantages are? Unless there's some advantage a damn sight
- > better than "it's easier for the user just to have one executable",
- > then this idea sucks.
-
- Agree.
- I not sure what a fat binary file means, but there is an interesting way of
- solving the problem with executables and processor architecures. By inventing a
- new instruction set (one that could easily be converted to 68K, PPC etc.) and
- having executables with that instruction format, one executable could be
- executed on any processor type provided there is some translation program
- available. This has been done in many systems, such as the TAOS operating
- system. (I suppose Java's ByteCode is a similar idea.) The conversion would
- only be needed once, when loading the executable from disk, and would hardly
- slow things down. If the instruction format is well designed, it would probably
- need less space on disk than native code and a lot less than two executables in
- one.
- I'm not sure if this is what the original poster suggested, but in the case of
- AmigaOS I think it would make a transition from the 68K to any other processor
- (not only PPC, if desirable) smooth.
-
- Anyone interested in more info on TAOS should read this article in PowerPC News:
- http://www.computerwire.com/powerpc/articles/news/1292.htm
-
- --
- Fredrik Hellgren
- fleck@ludd.luth.se
- http://www.ludd.luth.se/users/fleck/
-