Subj : Re: newbie questions about porting JS to embedded processor To : netscape.public.mozilla.jseng From : "Frederic Rudman" Date : Mon Dec 15 2003 03:42 pm WOW! Talk about a great review for SM! Thanks for taking the time to reply. Well, it seems I now have 2 good solutions to pick from (as per the posts on this newsgroup) namely Spidermonkey and Nombas' ScriptEase. I admit I'm leaning towards SM with 3 caveats (repeats of previous posts): 1) The doc in SM seems to be harder to work with than SE 2) The performance of SE seems to be somewhat better that SM (not a huge factor for us but we are going to use it in an embedded environment and it will be the prime UI engine driver for the device so I have to be careful). 3) In general, I don't mind paying for software: it supports the developers and it buys you some level of support. However I concur with Steven Cole's comments below, namely that the response on this newsgroup seems to be extremely responsive and helpful. Decisions, decisions, decisions... I'm going to noodle this over (and if anyone cares, I'll be happy to tell them the result!). In the meantime, any further comments or opinions are very welcome. Thanks, Frederic Rudman BTW: for those who have replied to me directly by email, I will reply to each one, of course, but I was away for a few days and I have to deal with a huge backlog of email (most of it spam, of course, but for which I have to move carefully lest I delete a good one by mistake). "Steven C. Cole" wrote in message news:brdf34$p251@ripley.netscape.com... > Brendan Eich wrote: > > More than a few have done so successfully, but I have lost contact with > > them (scole, formerly @planetweb.com, may know more). I'll let others > > followup here. > > News of my disassociation with Planetweb was a bit premature; still, my > involvement with Mozilla is now on a much more personal level than a > professional one, as before. > > Planetweb successfully used Spidermonkey on a number of embedded > systems. It ran (runs) on Mips CPUs, Intel CPUs, ARM CPUs, Hitachi > CPUs, and probably others I don't recall. Spidermonkey runs above our > porting layer and never required any modification moving from one CPU to > another. (Though we did turn on the 'NaN is broken' code for all our > CPUs, since that was easier than figuring out a test to see which worked > and which didn't.) > > Oh, and we had to replace a lot of the time stuff, because embedded OSes > often don't even contain the concept of time zone or daylight-saving > (which we provided instead as an application-level service). > > Memory space was never an issue, as Brendan's prior note shows. We > never clocked the engine --- we used it to support small scripts. And > the time spent running scripts was generally dominated by dom-level > stuff, not the scripting engine itself. > > If you're worried about the "cost" of using this engine, (i.e.: is it > really supported? Will you be able to get help if you need it?) stop > your worrying now. Spidermonkey support is amazingly responsive (and > has been for years). The documentation is lacking somewhat (mostly > because none of us volunteer to get it fixed), but the newsgroups are > full of very helpful people. I've never had support this good with any > commercial product. Primarily because you're actually talking with the > talent that can get things done here, instead of customer contact > personnel. There's truly something to be said about an absolutely open > development environment. > > --Steve Cole > .