Path: csiph.com!newsfeed.hal-mli.net!feeder3.hal-mli.net!newsfeed.hal-mli.net!feeder1.hal-mli.net!border3.nntp.dca.giganews.com!border1.nntp.dca.giganews.com!nntp.giganews.com!postnews.google.com!glegroupsg2000goo.googlegroups.com!not-for-mail From: Helmar Wodtke Newsgroups: comp.lang.forth Subject: Re: Which standard? Date: Sun, 25 Mar 2012 08:51:39 -0700 (PDT) Organization: http://groups.google.com Lines: 52 Message-ID: <17352675.207.1332690699889.JavaMail.geo-discussion-forums@ynuu20> References: <30794f7f396fa9ad7ec52f034288c527@dizum.com> NNTP-Posting-Host: 62.158.113.131 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable X-Trace: posting.google.com 1332690700 3778 127.0.0.1 (25 Mar 2012 15:51:40 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Sun, 25 Mar 2012 15:51:40 +0000 (UTC) In-Reply-To: <30794f7f396fa9ad7ec52f034288c527@dizum.com> Complaints-To: groups-abuse@google.com Injection-Info: glegroupsg2000goo.googlegroups.com; posting-host=62.158.113.131; posting-account=nibe3QoAAADcYL8fC0WC6vCas4By1Xgn User-Agent: G2/1.0 Xref: csiph.com comp.lang.forth:10452 Am Mittwoch, 21. M=E4rz 2012 15:13:50 UTC+1 schrieb Nomen Nescio: Hi, > Is there any consensus on which Forth Standard is a good starting point f= or > a new implementation? No. > I've done a little reading and I understand Chuck > Moore was not in favor of certain standards to the point he didn't consid= er > them Forth. He did a different thing than what you call "standard" as he understood For= th as a personal programming tool. > If no standard is a good starting point, is there any consensus > on what a reference implementation should be? Any pointers to doc, articl= es > etc. would be welcomed. Thank you. If you want to start, I think best is to know Forth and have a need to do s= ome implementation by yourself. Otherwise use something that exists. You have the choice from a lot of vend= ors - I think also the commercial ones give you free access to implementati= ons that are far more than enough to learn about Forth. If you really want to make an implementation by yourself - don't use ANS or= 200x at the first stage... Just ask yourself if you need it. But: keep your system close to ANS or 200x in a way that makes it possible = to convert your System into one that understands ANS. I implemented HelFORTH/forth4p and this system originates in a very nice "t= raditional" kind of implementation for the "lonely wolf": RetroForth. Now m= y system is near-complete ANS and implements some 200x if you start it with= out switches. It features a FFI and JIT-compilation. Why I came to ANS at the end? That is benchmarks and test suites. Especiall= y the test suites are something useful if you implement by yourself. It's also fantastic to get programs from people and see that it runs on you= r own system. But as there are 1000's of implementations out there you coul= d call "they are like Forth", people that share code use something commonly= known. At the moment it's ANS or at least things that are not that far awa= y. If you have no own idea what your Forth should be different to other ones, = don't develop an own one until you know Forth. If you know Forth, you have = an idea why (or why not) you want to implement your own system ;) -Helmar