bob_2 is correct as far as I know, but IMHO this kind of play with what are known to be normal spelling variations confuses people, and a computer programme = a computer program. Similar with analog(ue)—it is recognized that there are computing devices that operate by translating numbers into physical quantities, but why make an artificial distinction where
Why, if we do not have to? You can, e.g., buy Field-Programmable Analog(ue) Arrays. I could not tell you off the top of my head for which specialized applications their current performance would be superior to adapting the same computation to a FPGA.
[NB
“Analog computing” means that instead of signals just being discrete values (1 or 0 in a binary computer), they can be any of a continuum of values.
Wouldn’t we simply call that infinite-valued logic?
]
The old electronic analogue computers should not really be distinguished from what we term digital computers simply due to their operating with continuous values. This seems to be a retrofitting of terminology. There is no reason an analogue computer could not or would not deal with quantised or sampled values if it fit the domain name of what was modelled. Moreover implementation with sampling or quantisation would be perfectly reasonable.
The whole thing gets stretched to the point one really needs to be clear of what is an analogue.
One example is designing loudspeakers. In particular the bass driver and cabinet. The well understood theory (Thiele Small) analyses the system as composed of masses, compliances, and damping of air and speaker coupled to an electric motor. There is a well known direct electrical analogue to resistance capacitance and inductance. So much so that it is quite feasible to usefully model a loudspeaker design in this domain.
However a common trick was to use Spice to model this electrical system. So a digital model of an electrical system modelling a mechanical system.
You can argue that Spice isn’t an analogue - it performs some basic and modified nodal analysis. But it does perform time stepped transient analysis - which one could say is creating a numerical analogue of the real thing.
Any discrete event simulation might be considered as a numerical analogue. Whether you implement it in a general purpose computer or dedicated circuitry is a matter of convenience, not form.
There is now, but it was not always so. Up until the late 80s or so, the form ‘programme’ was still in use alongside ‘program’ in the UK; the latter form gradually took over though; same with disc/disk although that change completed a lot faster.
I’m not sure analogue will go the same way as it’s a concept that is bigger than its computing application.
Yeah, whether or not analogue computers can be general purpose is a matter of design. It might be harder to design them to be general purpose, and there might be no good reason to do that a lot of the time, but it’s not flat out impossible.
The distinction between the two types of device I described definitely exists, and I read somewhere that they were described by the two spellings that I gave, but even if those aren’t consistently-used terms for the devices, the two types of device still exist.
The devices that simulate specific systems are relatively rare, because it’s usually not practical to create a new device for every new problem, but there are a few of them, and they aren’t like the devices that just add continually-varying voltages or the like, which can be reprogrammed in much the same way as digital devices.
As far as international spelling variations taking hold in technical fields, another example would be flat, round objects. In the US, those are usually spelled “disk”, and in the UK, they’re usually “disc”, but a “compact disc” is spelled that way in both countries.
I’m no expert on photonics, but from what I understand, it has equivalent basic components to electronic digital computers, and so would perform all the same tasks. They’d just do the same tasks a bit faster.