Re: Re (long): A catty comment from a fan of HP calc's of the the '80s Message #12 Posted by Mike Mander on 26 Nov 2003, 10:25 p.m., in response to message #11 by GE (France)
I don't know what HP uses to develop operating software for their calculators, but I doubt that it has changed all that much in the last 10-15 years - someone please correct me here if I'm wrong. When you imagine the functionality packed into what, 2Mb of ROM?, surely routines are mostly hand-coded, or at the very least hand optimized, in assembly language - not the easiest thing to verify and debug given the relative complexity of the software routines. As well, given what the code will finally be running on, namely a 4Mb Saturn CPU or slightly faster emulated system on ARM, I'm not sure how ultra-fast modern computers would benefit the developers of HP calculators? Are there sophisticated software tools that can torture test and validate the built-in CAS functions and simulate actual actual real-world use for example?
With regards to the long lineage of the software base back to the 28C, I would have to agree partially on that. I'm sure many of the underlying routines, unrelated to CAS functions, graphing, advanced symbolic manipulations etc., have remained essentially unchanged since the 28C. Those routines should be solidly debugged and essentially perfect by now. However, based on my experience, I would have to say that those parts of the 49G/49g+ *are* more or less bug free now...
Yes probably less than 10% has changed from the 49G to the 49g+, but are there that many new bugs in the 49g+ that did not exist on the 49G? Not talking about new hardware issues such as bad keys or CPU low-power mode switches causing display flicker of course...
BTW, I am a Macintosh user and try hard to avoid using any Microsoft software, so I am not actually "...accustomed to Microsoft-class software 'quality'." ;-) (Not saying that Apple software doesn't have some problems of its own as well...)
Seriously though, I agree that in many cases, software "quality" is poor these days and I wish all developers would agree to take a deep breath, sit back and actually thoroughly test and troubleshoot their routines, and run extensive offical beta programs, before unleashing their software on the general public. Instead, they are all in a mad rush to add more features and beat each other to market - often at the expense of the end users. Most commercial software isn't hand-coded in assembly these days, can run on GHz plus processors as has access to plenty of RAM, so the severe constraints of operating in something like a 49G wouldn't apply for programs from Microsoft, Apple, Adobe etc.
It would certainly be nice if calculators were perfect and never crashed or needed a reboot, but once again, given the enormous functionality and the constraints of the hardware, I think HP developers are doing pretty darn well already.
Just for interest's sake, note that I have also experienced software crashes requiring a "reboot" in VCRs, microwave ovens, printers, scanners, ethernet hubs, hand-held GPS units and cameras, to name a few, and many of these devices are likely much simpler than a 49g+ from a software perspective. (well, maybe not my Garmin eTrex Vista GPS)
Anyway, enough babbling from me now, I've said what I wanted to say... ;-)
--
Mike Mander
|