Path: csiph.com!usenet.pasdenom.info!weretis.net!feeder1.news.weretis.net!news.swapon.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail From: Paul E Bennett Newsgroups: comp.lang.forth Subject: Re: PAF: a Portable Assembly Language Date: Sat, 24 Jan 2015 13:48 +0000 Organization: Hidecs Consultancy Lines: 80 Message-ID: References: <6fb13d8a-5d8e-44fe-a28e-a7d4a6e9eab6@googlegroups.com> <655f9282-2153-4739-9e33-90475a84e7cd@googlegroups.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7Bit X-Trace: individual.net tIfMCNtsUL1NjktrdEQKAAAaYUVBBkxH78zPv7q2ZmRemq5XJQ Cancel-Lock: sha1:KdETg61+7yu8/uFTWE3yb9WFBLQ= User-Agent: KNode/4.11.4 Xref: csiph.com comp.lang.forth:35992 hughaguilar96@gmail.com wrote: > On Friday, January 23, 2015 at 11:01:42 AM UTC-7, Aliaksei Chapyzhenka > wrote: >> Hi All, >> >> I have discovered great paper from Anton Ertl about Forth based Portable >> Assembly Language (PAF): >> >> http://www.complang.tuwien.ac.at/anton/euroforth/ef13/papers/ertl-paf.pdf >> >> I plan using it in the tool chain for my processor project, and was >> looking for more information about PAF. Does anybody can help me with: >> PAF wordlist? any open source PAF reference implementation? >> any other papers, threads, information sources? >> >> --Alexey > > This has already been discussed in this thread: > https://groups.google.com/forum/#!topic/comp.lang.forth/YVjtNQwSIH0%5B326-350-false%5D > Is that where you discovered it? Anton Ertl's paper isn't "great" --- I > read the abstract, but didn't consider the paper worth reading. > > Anton thinks that the register set size can be abstracted away, but this > isn't true, as that is the one aspect of the target processor that has to > be standardized. I don't think that Anton knows enough about assembly > language to be the person who designs this portable-assembler. > > The idea of a portable assembler is good, as this would allow Forth to be > quickly ported to any processor that fits the basic profile (there is not > much difference between the MIPS, ARM and ColdFire, for example) --- all > that would be needed would be to rewrite the portable-assembler, and then > the Forth system program would assemble for a new processor --- but the > target processors have to be similar in that they all have the minimum > number of registers and a reasonably orthogonal instruction set (there is > a lot of difference between the 6502 and the Z80, or between the 6812 and > the 8086, so a portable assembler wasn't realistic for them). > > This is what I said in the thread mentioned previously: > On Tuesday, January 20, 2015 at 9:40:31 PM UTC-7, hughag...@gmail.com > wrote: >> This is a good idea. This idea has been around since the 1980s though, so >> you shouldn't take credit for it. >> >> You can't abstract away the register set size --- you have to assume a >> certain number of registers. This is why the idea is feasible now, and >> wasn't feasible in the 1980s --- because you can assume 16 registers now >> (this leaves out the 32-bit x86 though, but it is obsolete anyway) --- in >> the old days, processors had very few registers, and juggling these few >> registers was different on every processor (the 6812 and the 8086, for >> example). >> >> Your PAF won't work on the MiniForth though --- it will only work on >> mainstream processors. >> >> Anyway, what I'm doing right now is targeting the 8-bit AVR. This is the >> smallest processor that I can imagine my Forth running on (16 registers >> with 3 that can be used as pointers) --- porting it to a larger more >> powerful processor should be easy (although going in the other direction, >> from a powerful processor to a more limited one, would be difficult). I >> don't actually consider the AVR to be very useful, because it is not >> powerful enough for most applications and it is well on its way to being >> obsolete --- I am just interested in the AVR as a proof-of-concept >> platform. There was an article in Forthwrite about the MAF and PAF implementations and maybe also mentioned in Forth Dimensions. The Forthwrite on-line resource didn't manage to get all the early paper versions scanned and on-line but I have many of the missing ones in my library. An email to Graham Dunbar might also be worthwhile. -- ******************************************************************** Paul E. Bennett IEng MIET..... Forth based HIDECS Consultancy............. Mob: +44 (0)7811-639972 Tel: +44 (0)1235-510979 Going Forth Safely ..... EBA. www.electric-boat-association.org.uk.. ********************************************************************