[Freedos-32-dev] RE: UDI alternatives
Status: Pre-Alpha
Brought to you by:
salvois
From: Roberto M. <RMa...@ar...> - 2001-01-25 13:34:08
|
On Thursday, January 25, 2001 12:45 PM, Salvo Isaja [SMTP:sa...@li...] wrote: Hi, I write on this after a long time just to show I'm still alive. > Hello guys, > after one more day of UDI specs reading, I'm even more convinced that if > we'll use UDI we'll never write FD32. Until now, the FD32 Survey doens't > show any predominant preference, hence we still have many alternatives. > > 1. Using a cut-down of UDI, implementing only the basic features that we > need for FD32. UDI allows us to do that by returning UDI_NOT_IMPLEMENTED (or > something similar) for not implemented services. Writing such an UDI > environment allows us to use UDI drivers and can be expanded later, but such > an approach assumes that we know UDI, and personally after reading many many > pages I'm still lost. But I've seen that a person has answered in the > Survey that he has read the whole specs and finds UDI easy, so I'm waiting > for illumination... I've not read UDI docs deeply. Maybe seeing some sources we can clarify ourselves? I agree on this, we can implement only a sub-minimal part o the UDI interface first, then extend it to the full version. > 2. Since UDI is not properly what Guideline #1 of our System Specification > says, we could discard UDI and develop our own internal interface, as simple > as possible. It could be interesting the possibility to realize the UDI > support in a later stage with an external module. After all our modules can > customize the system environment in any way possible (guidelines #3 and #4). I prefer the first option, see below. > My personal preference is the second option, but I don't now if/how this is > realizable. I think UDI is very important, because a serious drawback of DOS > systems today is the lack of drivers for modern devices (I think to sound > cards, for example, or USB devices AFAIK), but we have to keep FD32 simple. You are right, this is the question. A custom interface is easy to realize, but drivers are not. The BIG problem with FreeDOS are drivers, and if we can find already-made drivers this problem doesn't exist. I think that if we really want FD32 to succeed, we must use some sort of public interface for drivers. Really, I think that nobody will use our system if drivers are difficult to find (Linux had this problem for YEARS). Having this from the beginnig is: - a pain for us, we are going to do a big work on it - a paradise for our users (and for us in the future), when we can simply recompile an UDI driver to get the latest enhanced card working. Note also that for now we need drivers for: disks, keyboards, monitors. Maybe we can easily find generic (E)IDE, FD, AT/PS2 and text mode VGA drivers. UDI is the first idea I had, because it is working and supported by many hardware vendors, and I've heard of already-made UDI drivers for Linux. But we can use another interface, posed that there are enough drivers and we think that they will be in the future (as the Linux native interface). We can ask the UDI (or the Linux) community if they can help in some way. Ciao. |