In the absence of any firm statement from any RISC OS company, any comment from myself is pure speculation and should be treated as such.

Chris @ drobe.co.uk

The third was a certain STB developer, but their drivers didn't work last I heard (that was a while back though)

I just hope that when the dust settles all of the products being produced will conform to the same API otherwise it'll just cause more work for our already overstretched developers.

Now, if one company concentrates on producing hardware, one company concentrate on developing the OS and drivers, and another concentrates on software...

Long term the profit will be in supplying the USB drivers for peripherals.

Might it be possible to provide a common core of shared modules, so the proprietary stuff is only the model-specific bit, and hence much smaller? eg a common USB webcam driver module, with a few custom tweaks to configure it for each given model? Even with such streamlining, everybody's going to have to keep the day job, aren't they.

all you need is an agreed framework for those drivers to work within to ensure developers don't have to support multiple platforms. This is what is looking a little unlikely and is not good. Look at how long it took 3D accelerators to really take off under Windows until MS muscled in with DirectX.

How much does one have to charge to recoup ones investment? Not enough to afford carving up the market between competing suppliers, I'm sure. And I'll bet much of the USB code will be common across many drivers - so many wheels being reinvented for so few punters each.

It's true that the market isn't big enough to be "carved up". It's also true that many USB devices can be accessed through a common driver. However the more complex devices (eg printers) will require specialised driver support but I wouldn't expect companies to truly compete in this arena, ie I wouldn't expect each company to suddenly write drivers for Canon printers, I'd expect them to offer a variety of models.

Again this does "carve" up the market to some extent but it's not direct competition.

Might it be possible to provide a common core of shared modules, so the proprietary stuff is only the model-specific bit, and hence much smaller? eg a common USB webcam driver module, with a few custom tweaks to configure it for each given model? Even with such streamlining, everybody's going to have to keep the day job, aren't they.

Yes they are. It's also worth bearing in mind that income from USB drivers is still income. I doubt any company would be trying to make a living off them alone though.

H'mm - how do you get an agreed framework if peolple are competing - the past is littered with standards killed by commercial and technical wrangling. Is the M$ of the RISCOS desktop Pace or ROL? - no sign of an imposed freamework there? Lacking an M$, open standards are the general answer. Of course there has to be a free exchange of related information, and agreement on who coordinates and arbitrates on it all.

Having an agreed framework does not stifle competition - if anything it encourages it because it prevents one company creating a monopoly which they can exploit and prevent others from getting a foothold. Successful examples of this include linux distributions and the CORBA platform.

As for coordinating standards, for the desktop market it's the job of ROL - pure and simple. They're in charge of the OS so they make the standards (or leverage Paces), preferably after consulting the hardware developers. ROL should also be fostering communication between interested parties - it's not hard.

lol. why do I bother?

Because you actually care???

As for coordinating standards, for the desktop market it's the job of ROL - pure and simple. They're in charge of the OS so they make the standards (or leverage Paces), preferably after consulting the hardware developers. ROL should also be fostering communication between interested parties - it's not hard.

Absofruitly. That is what should be happening. But it isn't, is it? Why not (aside from the obvious berrations of Mr Middlebrain)?

Up to three companies have RISC OS USB podules in the pipeline, based on a number of names that I've been given. Two might be the same card with two companies working in concert though.

A fairly major RISC OS hardware player has a card working fully, to the extent that they tested a BT ADSL modem and found they hadn't got the /modem/ drivers working properly - implying that the USB side is finished.

AFAIK this is all about podule cards, not in-built USB in new hardware. Should we expect that there's a standard to tie the two types together, so writing device drivers will be easy and work on both types of interface? Some of the names involved seem mutually exclusive

So, should we expect announcements RSN...?

AFAIK, Castle is working on one card, too.

But that's not much more than rumours

So come 2003 it should arrive.

<windup>even open source projects usually do better than this. they tend not to hide things from eachother </oh not him again>

<blatant flamebait>in fact, if the RISCOS USB developers went open source, they could all leverage eachother's software talent and sell the profitable hardware - or is the story all wrong that software is the loss-leader that sells the hardware, and people are not really so dumb as to think they can make a profit out of selling USB drivers to eachother? </ducks for cover>

<ignoramus>Is anybody coordinating this kind of thing anymore?</nosey parker>

Doesn't seem like it does it?

<windup>even open source projects usually do better than this. they tend not to hide things from eachother </oh not him again>

Yes but they have coordinators

<blatant flamebait>in fact, if the RISCOS USB developers went open source, they could all leverage eachother's software talent and sell the profitable hardware - or is the story all wrong that software is the loss-leader that sells the hardware, and people are not really so dumb as to think they can make a profit out of selling USB drivers to eachother? </ducks for cover>

Except that the hardware isn't profitable because you can buy it elsewhere and cheaper than from a RISC OS dealer. Selling the drivers is the only way to make income

Except that the hardware isn't profitable because you can buy it elsewhere and cheaper than from a RISC OS dealer. Selling the drivers is the only way to make income

How many teams can the software market support, then? I know, why don't they cooperate, with a legal framework ensuring nobody can pull out and take their piece of the puzzle with them? I am sure that the margins on the manuals, bundled hardware and support deals would remain, and erosion due to lost royalties on the software would be far less than the exponential savings made in sharing the workload across a bigger market. Oh, that legal framework? Answers on a postcard please.

There, got what you wanted yet Lee?

Leaps for PC magazine and searches adverts for an Acorn compatible USB podule.

I take your point but once you've sold a few thousand of them then there are no more sales to make. Long term the profit will be in supplying the USB drivers for peripherals.

How many teams can the software market support, then? I know, why don't they cooperate, with a legal framework ensuring nobody can pull out and take their piece of the puzzle with them? I am sure that the margins on the manuals, bundled hardware and support deals would remain, and erosion due to lost royalties on the software would be far less than the exponential savings made in sharing the workload across a bigger market. Oh, that legal framework? Answers on a postcard please.

You don't to open source drivers though, all you need is an agreed framework for those drivers to work within to ensure developers don't have to support multiple platforms. This is what is looking a little unlikely and is not good. Look at how long it took 3D accelerators to really take off under Windows until MS muscled in with DirectX.

There, got what you wanted yet Lee?

No - I want to use USB from my RiscStation but it has no podule slots!!!!!