amiga-news ENGLISH VERSION
.
Links| Forum| Kommentare| News melden
.
Chat| Umfragen| Newsticker| Archiv
.

[Login] [Registrieren] [Passwort vergessen?]

< Nächste MeldungVorige Meldung >
19.Feb.2003
Christoph Gutjahr (ANF)


Interview mit Ben Hermans
Angesichts einiger Gerüchte bezüglich des Status von AmigaOS 4, die gerade im Umlauf sind, und der Ankündigung der Partnerschaft mit SciTech haben wir Ben Hermans (AmigaOS 4-Projektleiter, Managing Partner von Hyperion Entertainment) kontaktiert und ein kurzes Interview mit ihm geführt. Eine deutsche Übersetzung folgt in Kürze.

Amiga-News: There's a rumor that OS4 is further delayed because Amiga Inc. are unable to contribute money to the project. Does your business plan rely on financial contributions from Amiga Inc.?

Ben Hermans: No, it doesn't. It has to be recognized however that software development costs money and that more money means more development resources and hence a shorter development cycle.

Amiga-News: You have to pay three senior developers, and the only known source of income for Hyperion during the last 15 months was the work on the Teron/AmigaOne BIOS (and Quake 2 of course, but that probably made just enough money to buy Steffen a Pizza). What/who pays the wages of your developers?

Ben Hermans: Contract work as you pointed out. Our work for Mai is ongoing and has been quite extensive.

There always will be new chipsets to support, new CPU's to support, Linux related work to be done etc.

We just delivered a substantially upgraded firmware for the AmigaOne/Teron hardware. More news on this will follow.

During the OS4 development cycle we also ported two Mac games (Nobody Lives Forever and Gorky 17), which brought in revenue.

Amiga-News: Still, the idea that you'll have to acquire additional funds if OS4 gets delayed some more makes sense. Your only option to acquire more funds would be to assign senior developers to other projects. In other words, if you don't match your current dead-line, OS4 will get even further delayed because the Friedens and/or Steffen Häuser would have to work on some Mac-Ports or contract work not related to OS4. right?

Ben Hermans: Absolutely. In fact, I dare say that if we had been in a position to concentrate solely on OS 4 during the past 15 months, OS 4 might already have been finished.

Amiga-News: Could you explain the legal status of AmigaOS4 to us? You said you "own" it,and Amiga Inc. are entitled to buy it back at a (quote) "very reasonable price". Is Amiga Inc.'s right to buy the OS back transferable, i.e. could they sell it to a third party? What happens if Amiga go bankrupt, would you still be able to sell the OS as *Amiga*OS? Would the party that acquires Amiga Inc.'s assets also acquire the right to buy AmigaOS4?

Ben Hermans: Our contract with Amiga does indeed contain an IP buy-back clause, which means that Amiga can acquire the work done by us on OS 4.

Once Amiga has acquired the rights to OS 4, theoretically a third party might buy it from them but certainly not with a view to burry it and halt development. If no new version of OS 4 is released within 6 months after release of OS 4, our right to develop the OS further and indefinitely relives.

If Amiga goes bankrupt, our license is commuted to an exclusive license to develop AmigaOS further, for any platform, not just PPC. This legally excludes a potential new owner from exploiting the source-code himself.

There is therefore nothing to be gained from buying Amiga's assets in the event of bankruptcy unless the new owner is willing to work with the OS 4 development team in good faith.

Amiga-News: Could you try to describe the current status of OS4 without using phrases like "everything's right on track" and "99% finished"? Is the integration of the 68k emulator finished? Are there external beta testers running ExecSG on their system?

Ben Hermans: We are now down to 8 outstanding items of the feature-list we released some time ago.

Only one of these issues is a show-stopper in the sense that it could introduce further delays (the emulation integration), the remaining issues are minor, several of which will be taken care of still this month.

By way of example: InstallerNG, one of those 8 items I mentioned. It is not finished yet. This won't hold up release of OS 4 as we can always switch back to the old Commodore installer. Another example: one Reaction class still needs to be taken care of. We are confident this will happen still this month. There are more of these minor issues, 8 in all like I said.

We are still working on the emulation integration. Unfortunately the external developer working on it had several pressing personal and professional issues to take care of which resulted in some delay. This is one of those instances where a bigger budget would certainly have helped. We nonetheless hope to wrap this up very soon as the work is now in such a state that full-time Hyperion developers can step in and take over.

Exec SG unsurprisingly sports all the functionality listed in the feature-list as already implemented. Several developers have access to it and are running it on their Cyberstorm PPC and AmigaOne hardware in order to develop device drivers. Meanwhile a team of 77 beta-testers and translators have been testing and localizing OS 4 components since many months now.

Amiga-News: The Alt-WOA scheduled for the 26th of April has been cancelled recently, because of the fact that "OS4 may not be ready by then". Were you involved in that decision? What does "not ready" mean in this context: "not ready for sale" or "not ready for demonstration"?

Ben Hermans: The organizers demanded ironclad guarantees that OS 4-A1 combo would be on sale at that time.

As a matter of principle, I informed them that the only way we could guarantee that now, would be for OS 4 to be ready NOW. It is not and I therefore declined to give them this guarantee as a matter of principle.

Huge corporations like Microsoft (Windows 95 released in 96 and Windows XP ring a bell?) and Intel (Merced/Itanium anyone?) with budgets larger than the gross product of most third world countries have missed their deadlines as a matter of course. It is simply unrealistic to expect a 100% guarantee. Life doesn't work that way. Not for Microsoft, not for Intel or Sun and certainly not for a small company like Hyperion. Do I expect that OS 4 will be ready by that time? I most certainly do, based on the information available to me at this point. But it would be intellectually dishonest to claim that I can guarantee that 100%, I have been in the business too long now.

Amiga-News: What kind of development tools can we expect for OS 4?

Ben Hermans: Development for OS 4 will initially need to take place with a barebones GCC. We already have Linux based cross-compilers as well as an OS 3.x based cross-compiler.

Moreover, Olaf Barthel has been working hard on a completely AmigaOS native GCC implementation which does not require ixemul and instead uses his own C runtime library.

We also hope to convince the VBCC maintainers to produce an AmigaOS 4.x version.

Amiga-News: Your partnership with SciTech was a positive surprise to most users. But supporting 180 different graphics chipsets definitely sounds like overkill if you would be only targeting desktop users. Is this partnership a first step to make AmigaOS more interesting for "embedded" appliances, like kiosk systems or Infochannel services?

Ben Hermans: The partnership with SciTech has several very important advantages.

First of all, we are assured that we will always have access to 2D drivers for the latest graphics cards. SciTech's expertise in this area is unrivalled and the manpower they can bring to bear is unmatched. SciTech has built up close relations with many graphics card manufacturers over the years. They can get hold of the required chipset documentation much easier than anyone else. This partnership essentially frees us from the job of ever having to write 2D drivers again and instead allows us to redirect resources elsewhere.

Secondly, Hyperion and SciTech will work together on 3D drivers. SciTech recognizes our expertise in this area and combined with their know-how, chipset documentation and development resources we can be similarly cut down on development time for 3D drivers.

Thirdly, your point about embedded systems is very valid. You don't need well over 170 chipsets to be supported for a desktop OS although choice is always a good thing. In the embedded space however choice may well be crucial when it comes to deciding which embedded OS to choose. There are a number of chipset producers competing in this space which won't win any prices for performance but certainly for cost-effectiveness and low power-consumption. When you are dealing with embedded devices, you are dealing with high volume productions. It's not hard to understand that saving say 10 USD per unit on the graphics chipset translates into massive savings when you are dealing with a production run of several thousand devices. Likewise, some devices may well demand low power-consumption for various reasons (cramped housing, hot environment or simply energy conservation). For such devices too a wide choice of chipsets may well tilt the balance in favor of a specific embedded OS. Let me give one final example: kiosk and information systems. We know of well upwards of a thousand Amiga based systems still in use for this type of technology all over the world. These people are looking for a way to upgrade their aging systems but they are not satisfied with the TV Out quality produced by most current graphics cards except Matrox. By adopting SNAP, we are giving them that option.

Amiga-News: Some people were concerned about the way SciTech's SNAP technology gets integrated into AmigaOS. Will programmers have to deal with a completely new API, or even two independent APIs?

Ben Hermans: No, the idea is that SNAP operates solely as a display driver, a monitor driver for Picasso 96.

Amiga-News: Will SNAP display drivers be any slower than "native" AmigaOS drivers? SciTech seem to be using an OS independent binary format.

Ben Hermans: As some benchmarks on SciTech's site show, SNAP is fully accelerated and quite frequently outperforms the "official" drivers. We have no concerns whatsoever in this area.

Amiga-News: Given your tight resources, SNAP will certainly not be included with OS4.0, will it? When can we expect SciTech's technology to be available for AmigaOS4? What kind of display drivers will be included with AmigaOS4.0?

Ben Hermans: SNAP relies on the SciTech x86 Bios emulator. This is embedded in the AmigaOne firmware but not present in the Cyberstorm PPC for instance. We are still examining whether it makes sense to have the x86 Bios emulator present also for the BlizzardPPC or Cyberstorm PPC. This may well not be the case as it is very CPU intensive. Moreover, due to the very limited bandwidth on the Amiga classic, operating state of the art graphics cards (to the extent that they even exist in PCI form) can be considered overkill. Having said that, Forefront Technologies is still working on "native" P96 drivers for the Radeon range and they are making good progress.

We don't see any overriding reason why SNAP cannot be included with the AmigaOne version of AmigaOS 4.0. SciTech will migrate the codebase to Linux PPC (thus ridding us of any potential endian conflicts) and we will take it from there. This should be a fairly trivial job for a company which specializes in porting software from one architecture to another and from one CPU family to another.

Amiga-News: Is there anything you want to say in closing?

Ben Hermans: Yes, I want to thank everyone for their patience. I realize that it has been a long wait and that the wait is still not entirely over yet. On the other hand, I want to emphasize again the massive amount of work that was accomplished over these past 15 months. For the first time since the demise of Commodore, the entire AmigaOS source-code has been consolidated in one central CVS repository. A maintainer has been assigned for nearly every OS module. The old Commodore "bug and improvement suggestion" database is available to the OS developers. We have replacements for all still relevant Kickstart 3.1 modules and have implemented nearly all of the functionality outlined in the features document. No-one but Commodore has ever undertaken development on the AmigaOS on this scale. The result will be well worth the wait.

Amiga-News: Thanks for taking the time to answer our questions, Mr. Hermans. (ps)

[Meldung: 19. Feb. 2003, 23:25] [Kommentare: 58 - 24. Feb. 2003, 08:12]
[Per E-Mail versenden]  [Druck-Version]  [ASCII-Version]
< Nächste MeldungVorige Meldung >

.
Impressum | Datenschutzerklärung | Netiquette | Werbung | Kontakt
Copyright © 1998-2024 by amiga-news.de - alle Rechte vorbehalten.
.