Mega post. Part 3.

ZXNet echo conference «hardware.zx»

From Dmitry Malychev To All 12 October 2005

Hello Lethargeek Mega post. Part 3. Well, it came to “concrete proposals”, which I did NOT WANT to deal with. But has to. I wonder why respected hardware engineers reduce implementation problems first just to danger “hands-on”, that is, intervention with a soldering iron at the ready, at the very least, but working board? At the same time, apocalyptic pictures of “blue smoke” from “destroyed entrails" "now former" Spectrums. Actually, intervention with a soldering iron is personal matter everyone, "for 50 years now, manufacturers of electrical appliances and electronics have been writing in instructions: . And in magazines like "Do it yourself", they publish diagrams of all kinds finishing touches And because of this, the world did not slide into disaster and the abyss of heaven did not open up" - Conan. Why does the bells and whistles have to be done “from the inside”? It's not in vain that I said that the end user-coder should be deeply impressed by how brilliant (or on the contrary, what crustacean) way this is realized. He just does certain things agreements and that's it. Let us recall the early history of cloning in our country, when various models differed "primarily the way the video controller is implemented." Yes, just by quantity buildings in The clones of the 48 cars alone are worth a look! And nothing, almost everything is branded toysthey worked everywhere. And the incompatibility was due to port decryption, "untied" bus or INT signal. What has already appeared for software written for a specific clone, not under the branded Spec (that is, the coders have already used the nuances of a specific diagrams). So, why the possibility of external implementation by our ATM professionals not even being considered? Is it because on all ATM models, except 2++, the system connector missing? Or is it not planned in the future? Although a few pages ago we read: Splinter> Speck has already acquired something like an Audio card, I'm talking about GS, of course. According to my video Splinter> needed if ` and expand, then undoubtedly using a video adapter. Here it is only necessary determine what ` better: Some kind of video from the PC from the Stone Age, or put together absolutely individual ` device, as in the case of GS. CHRV> Yes, I absolutely agree with you, and I am working in exactly this direction. CHRV> As already ` said the v9990 chip will be studied (since we can get it and it’s good documented). CHRV> considering that most modern real estate developers already use CHRV> "modified", a Sometimes "used" Spectrums - what are we going to do with them? This is actually is"global problem", i.e. we (or you) cannot make it separate additional fee. Oppa! This means that v9990 can be made as an external video “card” (with GS compares), but something new is not only (I admit, to some) uninteresting, but also IMPOSSIBLE. How about v9990 going with Communicate with Spec - only through ports? Or will the memory still be displayed? What is required from an external card? Two ports for quarters. Memory 2*16 kb (or even 2*8?). Provide interception accesses to two “screen” memory pages. It's even better to just intercept records, Moreover, let the percent write to real pages (so that there is no later reading from video cards), and even if the “old” video signal is still generated - we same better, we can connect a second monitor. And of course the selection scheme (I emphasize not "mixing", but SELECTING) the next pixel from one or another screen... (yes By the way, mixing can be implemented at the same time - since you already have your own video memory - to get a “real” gigascreen, like No-flic in the emulator). The device will probably turn out simpler than General Sound ;) Or I (I fully admit) did not understand something and the system connector is not suitable for this is it good? So then we should not engage in empty talk about the dangers of “rule of arms”, but agree on a single standard for the most connectable connector, allowing connect the mostunimaginable devices, publish diagrams of modifications and/or adapters for main clones, work out agreements on potential bells and whistles (yes, this means a lot chatter, but not at all not idle chatter), and by the way, provide for the possibility of cascade connection devices, not conflicting resources, so that not everything is through tees. And the whole problem "hands-on" will disappear by itself. Then you will only need to “put your hands in” and risk your computer ONLY ONE TIME, at installation/modification of the connector. Or maybe I’m so illiterate, and all this has already been done? What then? dear pros are silent about it, and only scare with “blue haze” and torn guts computer?