XCore Exchange

Hi All,

I seem to remember reading a loooong time about about Chris Hinsley's TAOS operating system running on transputers.. I think it was in Edge magazine in the late 80's, early 90's. They had a GUI with some raytracing apps and what not. Anyone remember what became of that and, more importantly, if it ever got released?

Thinking about it, creating a micro-kernel with servers running on different cores, connected through links, and some external memory thrown in for data / app-space, an MP-64 could work nicely....

Yvo

Just google "taos transputer" and you get a whole bunch of info.
Don't think it ever took off.
Btw, I think putting an OS on the XMOS chips will just be a waste of their limited resources. I'm especially worried about the lack of memory management, which the transputers did have.
Can't remember if there ever was a proper distributed OS for the transputer. For instance, Helios was more a front-end OS running on a 68k with some back-end capabilities on the attached transputers.

I remember one of the Taos people visiting us and demoing it when I worked at BAe, Brough. It looked interesting, but the transputer was already on the way out. BAe still had one of my transputer systems that they'd bought a few years earlier, gathering dust in a cupboard. :D

Taos was pronounced Dow-os, because of the Chinese pronunciation of Tao. IIRC, it was supposed to have some of the properties of Tao.

I remember Helios. I tried it, but it was so flaky that it was unusable. The version I had actually ran standalone on the transputer. I knew one of the the people behind it as he'd previously set up Metacomco, developing compilers for the Sinclair QL, with which I'd been heavily involved. A friend of mine actually interfaced one of my transputer modules to the QL.

Leon

I remember contacting the TAOS people.

Their thrust felt very much to be one of taking a niche processor and creating an even more niche OS. I read a lot of hype and never actually saw product.

I admit it was an unheard of concept back then (or was it, share ware was very much alive) but had there been product it may have benefited the processor, and in turn itself, through an open source model.

I seem to recall the response I got to asking for a sample of their wares was very much, where was my penny.

An OS that makes a device less niche has been proven to promote sales. and aid longevity.

One of the Amiga Inc reincarnations tried to push an Amiga OS (AmigaDE) that was based on Taos / Elate. Unfortunately the Amiga community didn't believe in the project and many never gave it a chance, so I think the only fruits of the project were a couple of game bundles that would run on any PDA at that time.

I own the SDK and did some programming for it though. I liked the low level stuff, which IIRC is more or less Taos, but the graphical environment, Elate, lacked any kind of hardware acceleration and just felt bloated. This was on a PC though, could've been better on PDA.

johanar wrote:One of the Amiga Inc reincarnations tried to push an Amiga OS (AmigaDE) that was based on Taos / Elate. Unfortunately the Amiga community didn't believe in the project and many never gave it a chance, so I think the only fruits of the project were a couple of game bundles that would run on any PDA at that time.

I own the SDK and did some programming for it though. I liked the low level stuff, which IIRC is more or less Taos, but the graphical environment, Elate, lacked any kind of hardware acceleration and just felt bloated. This was on a PC though, could've been better on PDA.

It's been a running joke in the Amiga community that the only thing significant to come of all that was a snowman maker and a tip calculator.

The notion that an OS need consume vast ram and resource is insane.
The OS is just a program and should be as lean as possible. Unfortunately
the industry has bloated, i sugest a secure monitor w/fleshed out XC macro's.
Yes, we are going to bring memory to the core just not conventionally.

Thats my take,
Shawn