A few years back I dealt with a large company that produced a huge line of connectors. The catalog was about 2 inches thick. But for each different component they had both a part number and an order number, each with a lot of characters. And not a lot of ways to know if the pins would work in the housings you selected and would that housing mate with the other housing and would the sockets work with the pins.
I never did order any of their products, instead I went to a distributers catalog and ordered "plugs and matching sockets" that were adequate for the application, although possibly more expensive.
That is an other gripe, which is catalogs that provide no hint of pricing. Even a highly qualified price, dated, and subjecct to change without notice, and assorted discounts, would be valuable. Are hose 3 cent pins or 2 dollar pins? That sort of info would be quite useful.
Some product names are stupid, some are really good. Then, some names are well promoted and some are not. There isn't any inherent overlap in those four descriptors. The amount of money spent promoting it can have a huge impact.
For example, what engineer doesn't know exactly what a 555 is? The name has charm, memory and emotion associated with it. Step outside of the engineering world, however, and it's meaningless. It's not the fact that it's a number that does it in outside of the engineering world. Back in the day, much of the non-technical world knew exactly what a 486 was and that it was better than a 386. Who amongst that same set could tell you the meaning of Sandy bridge, Longhorn or Conroe?
If it's well promoted, almost any name can become meaningful.
I do know that the name of the game for many CE vendors -- until Apple came along -- for decades was to roll out as many variations as possible of essnetially the same product to meet different customers' tastes.
The problem with that is that you compete with yourself. Automobiles can be customized, and the buyer normally starts with a basic model and adds optional packages to get the car they want. Consumer electronics is less likely to work that way. If you have a range of different models, you must make your best guess on how many you think each will sell, and produce at least that many units of each model. You will invariably be wrong in some of your guesses, and the particular model a customer buys is also another model the customer didn't buy.
My immediate referent here is Sony in the glory days of the PDA market. How many different models of PDA did Sony produce in the Clie line? I lost track. And when Sony decided the PDA market wasn't profitable enough and corporate funds could be better allocated elsewhere, I wondered whether the proliferation of models, each with its own set of design and proiduction costs, was a contributing facter, because Sony couldn't sell enough of any particular model to see real returns.
Apple's notion of "make one model that satisfies 80% of the needs of 80% of the customers" has a lot to recommend it.
"But to defend Simon I think he was addressing product names for consumer electronics equipment rather than component identification for business-to-business sales."
Branding isn't just for end products. Part numbers are of course necessary for many reasons, but they don't preclude the branding and marketing of a family of components, even though component sales are B2B rather than B2C. Consumer awareness of and loyalty to a brand -- even a component brand -- can drive those B2B sales. If anyone doesn't think that branding of a family of integrated circuits, for example, is beneficial, I would ask are you as a consumer familiar with names like Pentium, Core, GeForce and Radeon? For those who answered yes, I would then ask, do you know the part number of the GPU in your expensive graphics card, or the part number of the microprocessor on your motherboard?
The problem with consumer electronics is that the products come and go obsolete at a fast and furious rate. With hundreds of models per year a company would be hard pressed to come up with memorable names. So just like the library book classification system, they turn to alpha-numeric codes which do follow logic, though not obvious to the average consumer. In any case, alpha numeric codes are still applied to memorably named products just to keep track of revisions.
Windows 8 may be a memorable product name, but there are many versions from home versions on up. Now there are also revised interface versions as well as touchscreen versions. And don't forget all the different service packs and updates needed to fully identify your Windows as it sits this second just before it gets changed by updating again. No way there could be a unique memorable product name for each.
I suppose they could follow the lead of drug companies, and invent words, but I doubt that would work. No sooner would they come up with a memorable name and it would be obsolete. Unless you are a company like Apple who wisely limits the number of products it produces, alpha-numeric product codes are the only way to go. Can you imagine assigning unique memorable names to semiconductors or I.C.s?
Now you're taking me back...the good old EL34 / 84 for audio,, and the classic ECC83 dual triode...
But point taken, Simon started this on consumer electronics. It's the marketing guys, not the engineers, who come up with the names.
I think maybe one of the reason's for Apple's success is that they stick to simple names. Iphone 4 or maybe 4S. But consumers themselves seem to like longer names and numbers - it's a status symbol. I am reminded of an old Philips VCR advert:
Salesman: "Can I help you, sir?"
Yobbo customer: "Yes, I need a new VCR."
S: "Well, this is our new model, one touch recording, on-screen menus for time delay recording, automatic rewind...."
YC: "Naaah....hasn't got enough knobs on it. What about this one?"
Blog Doing Math in FPGAs Tom Burke 2 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...