Path: csiph.com!x330-a1.tempe.blueboxinc.net!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: Dennis Ruffer Newsgroups: comp.lang.forth Subject: Re: CarrierIQ Software and Forth Date: Sun, 25 Dec 2011 12:58:29 -0800 (PST) Organization: http://groups.google.com Lines: 95 Message-ID: <760850.97.1324846709058.JavaMail.geo-discussion-forums@prez15> References: <78821811918436@frunobulax.edu> Reply-To: comp.lang.forth@googlegroups.com NNTP-Posting-Host: 108.213.68.47 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable X-Trace: posting.google.com 1324846817 19003 127.0.0.1 (25 Dec 2011 21:00:17 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Sun, 25 Dec 2011 21:00:17 +0000 (UTC) In-Reply-To: Complaints-To: groups-abuse@google.com Injection-Info: glegroupsg2000goo.googlegroups.com; posting-host=108.213.68.47; posting-account=kRenyAoAAACm35_ByGKKD3vBawNCQoka User-Agent: G2/1.0 X-Google-Web-Client: true Xref: x330-a1.tempe.blueboxinc.net comp.lang.forth:8348 On Saturday, December 24, 2011 5:19:56 PM UTC-8, Bernd Paysan wrote: > Marcel Hendrix wrote: > > 'Easily made run?' Are there cellphones with > >=20 > > 1. x86 32/64 bit hardware >=20 > No, but ARM hardware, and anybody who seriously wants to sell a Forth=20 > compiler in today's embedded system world already has an ARM target (VFX= =20 > Forth has an ARM target, SwiftForth has one, and Gforth runs on anything= =20 > that is a GCC target, anyways). >=20 > > 2. a provision to accept and run a user-supplied binary in a standard > > or > > open format (e.g. .exe or .dll) >=20 > Yes. Android is Linux. It took me a day to get Gforth running there,=20 > and most of that day was to improve the cross compilation support. Once= =20 > it did cross compile correctly, it worked out of the box on my Android=20 > phone. No debugging necessary, it just worked. I was impressed. >=20 > And for "the other cellphone": iOS is FreeBSD. I would not develop for= =20 > iOS, due to Apples knee-jerk policy about free software and especially=20 > interpreters, which apparently they forget completely when they install= =20 > spyware (Carrier IQ was part of iOS until version 5). >=20 > > 3. open and documented APIs (e.g. C-runtime library, SDL graphics and > > of course > > phone-specific stuff)? >=20 > Yes. Android is Linux, it is well documented, it has a libc (though not= =20 > glibc), it has OpenGL ES (not SDL), and even though the "official" way=20 > to program it is Dalvik, Dalvik uses C libraries for all the grunt work. = =20 > iOS's lowlevel parts are almost the same, though on top certainly there= =20 > is Carbon and Objective-C. >=20 > > A specific pointer would be appreciated. Given the Windows 8 hoopla > > I was expecting something to *become* possible, but searching for this > > kind of technical information is extremely tedious. (Especially when > > one isn't really interested in application smart-phone stuff). >=20 > Oh yes, it took me just a day to port Gforth, but it took me over a week= =20 > to find all the lose ends of the strings to pull to know where to look. = =20 > The Android NDK comes with examples which don't compile, and all the=20 > stuff that directs you towards Dalvik is misleading. You don't want to= =20 > go through Dalvik if you port a Forth to Android. You want a bog- > standard Linux libc program, you want to run it in a terminal session=20 > (androidterm/ssh/telnet), and thus you want the NDK, and to get a normal= =20 > cross compiler out of the NDK is another step, which is not explained=20 > properly. Look into gforth/arch/arm/android/config.sh (from the CVS=20 > repository) for more informations, the first few lines (with comments)=20 > describe how to generate a "feels normal" GCC cross compiler suite. >=20 > To make a more complete system than just a terminal Forth (i.e. one you= =20 > can develop programs the user interfaces with), you probably have to=20 > take androidterm (it's open source, it's a rather small program), and=20 > extend it so that all the stuff you absolutely need Dalvik for are done= =20 > there (like the interface to the keyboard, maybe some way to create=20 > Dalvik GUIs from the Forth system). >=20 > --=20 > Bernd Paysan > "If you want it done right, you have to do it yourself" > http://bernd-paysan.de/ Bernd, did you check your research into the gforth cvs? If this information that CarrierIQ used Forth gets noticed, it would be opp= ortune to be able to point people to something that could be used as a step= ping stone. It might even give us some relevancy again in that world. I even ran into a case where this would have been useful this last week. I= was at a factory in Shanghai and they ran into a slow eMMC issue. It woul= d have been easy to demonstrate the problem in Forth, but that is no longer= my job. The economics have changed to the ODM being responsible for deliv= ering working units and they decide how best to accomplish that. I simply = make sure that they do it right and give them clues on how to do it better.= Forth would be a great clue, if I had something to point at. An Android apk would be even better, but one step at a time. DaR