| > I intrigued by the support for the 28S these days. I still have my 28S (with working battery door) from when I was at University. But, while I was impressed by it's potential, I never loved it as a device to use I came to the higher end HP calculators with the 48SX, which was essentially v2.0 of the 28S. Aside from the 48's larger display and single keyboard, the two machines operated essentially the same, and my experience tends to align with yours. (Even if the 48 addresses the specific concerns you mention about direct access to common features, bidirectional I/O, and the split keyboard.) For me, the I way I'd put it is that these machines were better pocket computers than they were pocket calculators. Everything was just a touch too slow and overwrought to be as quick as you want a pocketable tool to be. In HP's defense, these machines were designed in the mid-1980's as a sort of next generation calculator architecture. They developed a higher level programming language (RPL) and a suite of tools for the purpose of getting themselves a solid foundation for continued work after the HP41 series (which was pushing the limits of HP's traditional SW/HW processes). So the The first RPL calculator in this new series was a business model (19B) very similar to the slight later 28C and 28S scientifics. Notably the 28S was sold in parallel with the 41CX, largely, I think, because HP recognized the appeal of the 41 and the fact that the 28 wasn't in a place to be a full replacement. The 48 was the combination of the two lines, learned from the issues of the 28, and was intended to replace both the 28 and 41 series. So like I mentioned above, the 28 really the first try at a high end scientific based on RPL and the 48 is the second. And for better or for worse, this is where funding for further development became highly marginal. There was a 48GX that doubled the clock rate and bundled the formula library, but my understanding is that that was really the last major development of the platform by the original engineering group, and they didn't ever really do a third version. (Which is probably where they'd have gotten it fully right, based on experience with other HW/SW products over the years.) One place people might argue with this narrative is that there were, in fact, two follow ons in the form of the HP49 and HP50. These bundled in additional open source libraries developed by the community (MetaKernal, IIRC) and at least one used an ARM CPU running emulation of the old HP Saturn CPU used previously. Neither were built to the same mechanical standards of the previous models, and I don't believe there was much continuity in the engineering teams. (The follow on work was done at least in part by members of the open source community that HP hired to briefly form a new calculator engineering group.) > and I've stuck with my older FX-7000G as the calculator on my desk. I use an emulated HP42S. |
The 28 and 48 series sat in a kind of awkward spot between pocket calculators and "real" computers. The programming language is essentially Lisp turned backwards, but for any task where you actually need that capacity, it must have been easier, even at the time, to just use a PC (in the wider sense, not just IBM PC and clones). The 48 had better communications facilities, so I imagine it might have been useful where pocket carry was actually necessary (it seems to have been popular with surveyors, for example, and I have seen them hooked up with a cable to a total station).