Date: Mon, 29 Jan 1996 19:50:43 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: nate@sri.MT.net (Nate Williams) Cc: terry@lambert.org, hackers@FreeBSD.org Subject: Re: Multi-Port Async Cards Message-ID: <199601300250.TAA05385@phaeton.artisoft.com> In-Reply-To: <199601292306.QAA09994@rocky.sri.MT.net> from "Nate Williams" at Jan 29, 96 04:06:41 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > Companies like Diamond, whose reluctance to provide programming info > > stems from their hiring of EE's to write their ROM code ... > ... > > > Maybe they will hire a software engineer before it is too late; maybe > > not. > > Are you saying that EE's are not software engineers, because I take > offense at that. Some of the best software engineers I know have EE > degrees (and don't have CS degrees BTW). Some of the best ones I know are HE or Solid State Physicists. 8-). The problem is not the ability to solve problems (which I think comes from teaching critical thinking skills and so all hard science types have an edge). The problem *is* that for a hardware company, software is always second fiddle to the hardware, and is used to "fix the hardware" instead of being written to be software. That leads to a tendency to take the first thing that seems to work as final product. Software, unlike Hardware, is not necessarily good just because it survives burn-in. 8-). A *good* video BIOS would work in protected mode or real mode and not care. If the programmer was truly a weenie, he'd give you an alternate entry point for all INT 10 calls that would work in protected mode as well as the ones that wouldn't so he could save his push/push/pop/pop in the "real mode" case. A *good* video BIOS would have a JMP followed by a structure that has a version number (in case) and is otherwise expected to be the same on all cards so the protected mode driver doesn't have to be rewritten for each release. A *good* video BIOS waits for veritcal blank instead of disabling interrupts to get rid of "sparklies" on non-dual-ported RAM access (cv: some ATI and Paradise VGA cards). Disabling interrupts might be OK if all you are doing is video I/O, but it sucks if you are using the console on a 100Mb/S router, etc.. Anyway, you get the idea -- I don't need to redesign Diamond's BIOS for them here (they need to hire someone who understands modular design principles to do it, since the guy who did it last time screwed up). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601300250.TAA05385>