Stephan Guillard interview (like the previous interview translation this was done with a combination of internet AI sites and my fixing some obvious mistakes. But I sure missed a bunch and some I'm not sure what he's saying as I am not familiar with French "tech" expressions. or everyday idioms, for that matter. It's been about 100 years since i actually spoke french):
There are only two French developers in the team of development of the OS4. Here a discussion with one of them, Stephan Guillard, who says some to us a little more on him and AmigaOS 4.
# could you present yourself at our readers?
In preamble: I give a whole from personal points of view below, therefore in a generic way, to add "my point of view" at the beginning of each paragraph:)
I am called Stephan Guillard, I live in the Lyons West. I am 38 years old, I am married and I have two adorable children (a girl and a boy).
In the true life, I work in a company which conceives and manufactures "testing machines" allowing our industrial customers to test their products, is in phase of R&D (when they seek the best design of the product to reach the performance aimed to lower cost), that is to say in phase of production (when they want to check that their product reaches the sold performance well, before delivering it). To give a concrete turn to this definition, our customers manufacture for example control units avionics, and want to check a few hundreds of parameters before delivering them to the manufacturer of plane... what is reassuring!
I created this company in 1999 with a dozen other engineers, like a "spin-off" of the Steria group (in other words, Steria is shareholder for a half, us for the other), one of the large European software firms in which I had passed ten years, according to the ordinary course in this type of companies. Now, my everyday job consists in meeting the customers Européans and American identified like "having a project" by our commercial, and building for them a solution, while being pressed on all the facets of our trade of engineering (data-processing, of course, it constitutes the "brain" of our machines, but also electronic, mechanical, hydraulic, acoustic, short all that it is necessary to put around the product of the customer to test it under realistic conditions).
Apart from this trade which impassions me, I have a life of family "normal"(*), and leisures common like the music, the cinema, the reading, the ski, the photograph, the voyages, etc.
(*): until 22h30. Afterwards, it is code in my office up to 2 or 3 hours of the morning, when I do not occupy myself of our American customers, time zones oblige :)
# When did you begin in the data-processing world?
My mother is math student, my father is an electronics engineer. At the house, one spoke less about the cows or the modern painters that of logic and science... though!
My father was one of the craftsmen of the micro French electronics; its laboratories the calculators left piloting of ARIANE and of some satellites, for example... they was the premises of what today is called Thales or Sextant avionics.
When I was around 10 years old (about 1975), it sometimes happened that I accompany him to the laboratory on the weekend (he worked 35 hours per day), when it was not the laboratory which came to the house... I was literally fascinated by the programmable machines (Microphone Nova, Kontron etc.) who were used for the development and the test of these space electronic charts or avionics. There the engineers let to me touch, they were then machines with consoles teleprinters, with punch cards etc but there was already something of magic in the fact of seeing being carried out a "complex" process defined by holes on charts... and also of odd in the fact that the least error of hole has catastrophic consequences.
I as frequently had right to the military samples of components of the time, right come from the USA, as the founders forwarded to the laboratory for evaluation, with docs, and sometimes complete system of evaluation.
Then, indicator which electronics interested me, my father was put to bring back its frequent voyages for me to the USA (and in particular to California, where its suppliers of components were, and its customers airframe manufacturers as Lockheed) of the objects which it was then impossible to find in France... They were, wholesale in the order if my memory does not betray me (and if my DMA interns bugge not, hein):
A computer which could make the 4 operations. At the time they had not found the means yet of making so that it goes back to "zero" to the powering, it thus posted anything, and it was necessary to withdraw up to zero...
A chart of development for 6502, the Kim-1 of Synertek. Ah... 256 bytes of RAM, 4 bill-posters 7 segments, a hexadecimal keyboard... My first love! -) (it is not MOSTECH, swallowed later by Commodore, which invented the 6502 that one finds in APPLE 2 and C64 etc, but Synertek, other start-up Californian, it even swallowed by MOSTECH:). This chart arrived with a spec. from the 6502 written... with the hand! As much to say that when I had APPLE 2, I did not need assembler to seize the computer code into hexadecimal. I remember still today, almost 30 years later, the opcodes kind LDA (#$$A9>), STA ($8D)...
APPLE 2. The most succeessfull machine which I saw to date. In 1978, 6 years before IBM PC, this machine had slots and charts of expansion (video, floppy, parallel, series, then later Z-80/CP-M, mouse, hard disk etc), a food with cutting without ventilator, an interpreter in ROM, an operating system, a very neat case, etc... and over the whole, a whole of handbooks describing to the last bit how the machine worked, how to program it, how to design charts of expansion, which largely explains the world explosion of the sales until the middle of the Eighties: before tidal wave PC, one saw of APPLE 2 everywhere. On the other hand, it was a very expensive object, and a configuration with a disk drive and a monochromic screen 80 columns APPLE could cost the price of a small car! To give idea of it to what the other micro computers of the time resembled (and to better include/understand why APPLE écrabouillés), look at the film Wargames, David has in its room "better" claiming competitor of APPLE 2, a IMSAI 80... I let to you judge.
This machine throne still on my desk, it boots on a hard disk controller of my design, with a OS also made house, and I make use of it every evening, inter alia like terminal of debug for my AmigaOne of development.
A TI-57, then a TI-58. One overclocks them by crayonnant with carbon the tracks of the physical circuit the clock, until they calculate false, which was not always easy to detect...
Then, by a stroke of luck, opened in the bottom of my home a shop of micro data processing (certainly the first of the area), which sold microphones of the time at SME, with made software of management house. Excel did not always reign on the market, before there was LOTUS 123, before it was Multiplan, before it was Visicalc (on APPLE 2) or... nothing! The managers of this store allowed me to pass there a good part of my spare time, in exchange of what some friends and me created some utility utility programs in BASIC to them, as well as a robot controlled by Sharp MZ80, who served the coffee and put sugar, the whole trônant in the window!
Arriving at the college, I fell on a teacher from maths who was insane furious: he had bought a TRS-80, and discovered in what the programmability of the machine was the ideal pen for its discipline. One not passed almost all the days the hours of meal to the canteen, but on his machine to make utilities of layout of functions etc. The fact of reversing the prof/élève report/ratio with 12h00 and of giving it in the good direction two hours later gave an odd impression.
Then, I had other machines 8 bit (Oric-1 etc), the middle of the Eighties was one incredible time for the microphone: not a week without a new English manufacturer not leaving a new machine (Sinclair, Dragon, DAI, Amstrad, etc). They all resembled each other in the broad outline (64k, 16 colors, cassettes with floppies in inaccessible option, its AY3-8912 etc), but each one had its personality and its aficionados.
In what relates to me, the virus was already irréversiblement established.
I preserved all these machines, in perfect operating condition, with their docs and software of time (yes, I know, I have the chance to have place, and especially a tolerant woman!). It is amusing to note that all my diskettes of APPLE 2 of the end of the Seventies or beginning of the Eighties work still impeccably, such as for example my original of Visicalc, a diskette 13 sectors whose label is written hand of Mr. Bricklin (the author), whereas the majority of diskettes HD of PC or even first CD-Rom which I used are perfectly unusable today... Always more bits, always less reliability.
Some other interesting machines were added to my collection since: NeXTCube with its charts DSP and its OS which was already 70% of MacOS X of today (the core Mach, slips by Objective C, it browser, the Postscript Display, etc), Silicon Graphics with its Irix which was the 30% remainders (the entirely vectorial office scalable, the multi-media one, etc).
The last in date is MacMini. The return of Steve Jobs to the levers brings back APPLE in the way of the machines which make dream, time forgotten since its departure towards NeXT.
# When and how you discovered Amiga?
One day of 1984 or 1985 (I tell of memory, it will be well somebody to contradict my chronology, it do not matter), I passed in front of the window of a store in which trônait A1000, surmounted of a screen 1081 which posted a scrollant colour gradation, which was difficult to believe for somebody who had never had more than the 8 basic colors and their half tones...
The machine was marked "Amiga". The program, it was a version alpha of the utility of diskette copy "Marauder". I remember as if it were yesterday, of the incredible animation of floppies etc.
I returned in the shop, and I met a band of brilliant eccentrics who "had seen the light", the band of Computer Shop Janal (Valence, Lyon, Chambéry, Saint Etienne, etc).
I had to discuss with Jacques Durand at least 8 hours this day there. I met the others thereafter (Ilan Sebbagh, Christian Bugnon, George Gaillard, Belkacem Benedris and others) and I make a point here of paying an immense homage to them, because they were animated by passion and not by the mercenary attitude, and they made it possible some generations of young people to discover data processing, and for more one, to make of it a trade which enables them today to nourish their family...
I owe to these friends as much as with my parents in the construction of the "data-processing" facet of my culture.
I thus learned that Ilan had fallen in stop in front of this A1000 which was shown in preview in a Parisian living room of data processing, and had literally torn off it with the distributor (Janal was Commodore distributor since the end of the Seventies, those which knew Procep will have a moved memory of it...) Good, Janal had already torn off not badly from prototypes of machines to the importer, of CBM PET2001 with the CBM 8000, but they were the best promoters of these machines.
While returning at home, I argued "sufficiently" with my parents so that this machine with the long course finds its place on the table of the living room! Mmh... More than 15.000 franks for a machine with 256 KB of RAM, a screen NTSC, a keyboard QWERTY, a chipset proto NTSC which did not even make a HAM, a pre-1.0 system, etc, it was inflated of my share. Especially that one could not even make use of it... without connecting it on a transformer 110/220!
But good. A true microphone 16/32 bit, that put on a snap. To launch 5 "CLI" and to make them work in parallel, it was incredible, absolutely inconceivable for any other machine with personal vocation of the time.
Like APPLE 2, the machine arrived with all its diagrams, and later one could buy a whole of docs of programming conceived very well (the famous ROM Kernel Reference Manuals, alias RKRM, expensive them also...), which were complémentés by the series of the books Micro Application devoted to Amiga which came then, some was extremely well done.
Here is how I discovered Amiga.
Around Janal, revolved (and still revolves) a population of people mixing developers, users with hobbystic vocation, and users professional, and quite naturally installation a relation between these people. I belonged to the first category, and I had the chance that some of the third agreed to pay me for the development of some tools, which enabled me whereas I was student, to buy up to date machines to me (successively A2000, then A2500/20, then /30 etc), as well as the ad hoc development tools (the long series of the compilers Manx Aztec/C, hold, the same one as I used on APPLE 2! As well as other languages, which corresponded in addition so that one asked me to work in my studies, Lisp, the Pascal, Prolog, etc).
By way of example, here some original applications which I at the time developed around Amiga:
A stereoscopic software of visualization allowing a doctor operating by endoscopy, to position in space 3d of the "standard objects" to evaluate the distance between two bodies. The principle was based on the creation of two sights shifted according to ocular axes', in synchronization with the angle and the x-ray one of the two endoscopic cameras used. In other words, I used the capacity of Amiga to change the contents of the screen to each image, alternatively to visualize the sight "eye right" then the sight "eye left" of the objects spread out in 3d. This vision 3d was then shown to the surgeon in superposition (genlock) with the respective images of two cameras droite/gauche remote-controlled inside the body of the patient, through stereoscopic glasses LCD, alternatively opaques/translucides on the left and on the right. At the time, except electronics to measure, impossible to carry out that with another thing that Amiga.
A system of medical data acquisition ambulatory nonintrusive. We designed and carried out portable Amiga (very reduced), with one 68000, 512K of static storage, a screen LCD etc, which made turn our very reduced version of Exec, and which acquerrait the signals resulting from 16 ports standardized on which one could connect a measurement of saturation of oxygen blood (by transparency rouge/infrarouge with the manner of the sensor of the Cauet Method), of the sensors of electrocardiogram, the sensors of flow and nasal temperature etc. This "portable" (1 kg with its battery) was connected to A2000 by parallel port, entirely parameterized in terms of configuration of the sensors, goes back and hour to beginning of acquisition etc then disconnected, installed in a holster carried by the patient, connected to the sensors. The patient left, then returned once finished acquisition. The memory of the portable was discharged in A2000, and all the acquired signals analyzed by a mountain of lines of code C, dedicated to the reconstitution of an electrocardiogram, etc. This system was used by the first laboratories being interested in the disturbances of the sleep (they sent the patients on their premises so that the fact of sleeping in a bed of hospital is not an awkward disturbance), or even for analyses of sporting behavior etc. When it east is a question of reaching the industrial stage, I discovered that the adventure of company was not that a business of technique... and very stopped there.
I also gained a little money at that time, by selling synthesized images raytracées by Sculpt3D, for example with manufacturers of pressure cookers, or standard lamps. This software was one of my most impressive revelations besides, so much so that by homage to his author, I entirely rewrote it thereafter.
Lastly, I spent a few free days of student's obligations behind the counter of one or the other of the Janal shops... as of innumerable nights to recover for the Janalesque community the contents of the diskettes Fred Fish with my modem 300 baud connected to waiter X25 of the campus of Doua :).
did # Comment on evolve/move your equipment and which is your current configuration?
In fact, after the end of the adventure of the system of medical acquisition, on which I had passed the majority of my nights of student during 3 years, I garaged Amiga at the cellar during... a long time. The 4000 had just left, in so far as I remember some.
I "made of Linux" during a good moment, on PC (being like everyone puffed up by the exponential progression of the performance/prix report/ratio as announced by Moore).
I thus have whole witnessed the end of Commodore, neither to the tumultuous episodes which followed myself from there, nor with the conflicts of people born at this time and still perdurant aujourd' today.
I réintéressé myself in Amiga one day of 2001 (or 2000? or 2002?) when Christian Bugnon more or less took me along of force to a living room to Cologne.
I fell on the stand Haage & Partner, on Sam Jordan and one of the Frieden brothers, discussing between them optimizations for WipeOut 2097... I looked at the play turning on one 4000, I had sorrow to believe that it was possible. I then discovered a posteriori that of which all those which had not left the boat as me laid out: charts PPC, video charts 3d, Ethernet charts, and on the software level, all that it was necessary to benefit from all that. There was also another member of Haage & Partner present (Markus Nerding?), who made me a demonstration of StormC 3. I found very sympathetic to finally be able to have an acceptable IDE (with my taste), integrated GoldED once.
I exhumed my 2000 equipped best, I bought in the USA a chart GVP-m 2060 with 128 Mo of RAM and a disc SCSI, I stuck to it Picasso 4 and one Ariadne 2 (nicely given by Janal), and I replongé (Linux less and less interesting me with the explosion of the deliberately incompatible distributions pushed by the mercenary attitude).
It is at that time that I developed a Linux/m68k pilot for chip SCSI of chart GVP, then AmiVNC, then some other gadgets.
I then had the occasion to repurchase one 4000 with a young engineer who sold very to create his startup Internet (it would have better done to keep his machine:), then the chart CyberStormPPC de Raphael Barthel, which sold it to finance its future "chart G4 de Phase 5".
I had just finished the first version Windows/Direct 3d usable of my clone of the Sculpt3D modeller, and a friend (David Larochette) launched me an amusing challenge: "you will never make Amiga version of this trick". Good. Rather than to rewrite very for API Amiga, I wrappé all the calls of Windows function towards calls Amiga, management of the windows with the 3d, and here are, a few weeks later, that turned.
What I liked as well for AmiVNC as for Skulpt/W3D, it was of "aller at the bottom of the things" on the corresponding fields: the programming network then the programming 3d. Our machines still allow that, whereas under Windows, it became differently more complicated...
It was also my motivation when I had created my hard disk controller for APPLE 2 (and associated software): to include/understand in an exhaustive way how the IDE functions.
Then, I bought Amithlon and I installed it on a PC at the house, "to see" (and to support H&P, to which I also bought StormC 4). I was allured by the speed of Amithlon, the code 68k constituting our OS 3.x and our applications turning several orders of magnitude more quickly on the JIT of Amithlon than on Nos 68060. What made me feel the nonsense not to have this OS in native version PPC.
On the other hand, I found odd that one cannot use the material 3d under Amithlon, whereas all the remainder was usable (network, hard disk etc).
I contacted the Frieden brothers to see whether they would agree to give me the files includes and docs necessary to the creation of one made software 68k, which they accepted, provided I refer about it to them to decide to distribute it or not once finished.
I made this code. But the incredible conflict battle field that became Amithlon meanwhile made impossible the payment of a Warp3D licence by the team Amithlon (Meyer, Frank, H&P), since it literally exploded. I understood at that time that Amithlon had any future, neither as a product, nor for the OS which it makes turn (OS 3.x) which would not evolve/move any more.
Consequently, Frieden required of me not to distribute neither the sources of the swrender (which would have made it possible to recreate a version 68k) nor binary the 68k. I thus put only on line the binary PPC WarpOS.
# What did you make with your Amiga since you has one of them, as much in term of programmation/creation that "additional" occupations, except OS 4.0?
In terms of play: I am absolutely not a player. Moreover, I have also PlayStation 2, but I only make use of it to develop code above:).
# You form part of the team of the OS 4.0. How did you enter the team, and of what you occupied yourself until now?
I did not take part at all in (nor interested by) the initial genesis of OS4. I believe to remember (without to be sure for it whole) that one day my friend Benjamin Yoris announced me that Frieden sought to constitute a team for this project.
On the basis of the principle that OS 3.x was died bus H&P would not develop anything beyond the 3.9, and that Amithlon did not have future, this option of bearing in PowerPC, with the prospect to remake in better certain not very satisfactory aspects, allured me.
I contacted Frieden, indeed they reached the stage of the conceptual reflexion, and were interested by my participation. Here, I formed part of the team, with the initial intention to work on the successor of Warp3D, "Nova".
For information, the team constituted itself completely at that time there, and did not move since. We are 43 developers distributed in the whole world, it never had there no important conflict between us on some subject that it is. This is largely due to the fact that we do not walk ourselves on the feet, nature having made that the sensitivity or the centers of interest of each one constitutes a complete panel to make OS4.
We communicate daily by a mailing list of developers, then another for beta testers (which are approximately 150). We have a waiter CVS containing the current state and all the former versions of all the modules of the system. We have also two waiters ftp (for the developers, for the betatesteurs), a tool of follow-up of the carryforwards of bugs (BugZilla, which rots you of malls until you corrected your bug) and Wiki allowing each one to write all that does not return within the framework of the other channels of communication.
About the development tools, formerly, to entirely compile the OS 3.x, the rumour, not far away from the truth, said that one did not have less than 9 compilers or assemblers turning on 5 different platforms. Today, OS4 can be compiled completely under OS4 by using only one compiler: GCC. One can also manufacture OS4 since other platforms by using what we call of the cross-country races compilers, available under Windows, Mac OS, Solaris, etc. Personally, I develop under Windows, by using Microsoft Visual Studio (my environment of preferred development), in which I completely integrated the includes and compilers OS4, but also PS/2 and APPLE 2, which enables me with only one tool to maintain all my projects for these machines, in complement of my Windows projects.
In terms of personal contribution to OS4, I was interested successively in several things:
Initially, to pose the concepts around the use of the MMU of PowerPC. How to use the Registers Segment, BEATS them, etc.
Then during a few times, with large step thing in fact (being largely mobilized by my projects hobbystic on PS/2, which was concluded by one second place in Evoke 2003 in Germany, behind Haujobb). During this time, Frieden coded tripotée of functions which constitute Exec/SG (since old Exec was 100% assembler 68K, one needed in any event all of rewriting).
Then when this "kernel" was put at booter on Amiga Classic/PPC, I helped with débugger here and there.
Then when first AmigaOne arrived, I made the error to raise the finger when somebody said "which one of them in the room know the IDE/ATA ?"... I have passed time very important to develop pilots IDE for OS4, inter alia because of problem of interaction between VIA and chip Ethernet at the time of transactions Ultra DMA, and owing to the fact that my machine of development at the time was equipped with a barette of RAM which flippait a bit in a random way all "N" giga bytes, which I put of the months to identify for what it was really (a problem of RAM).
Initially, I made a pilot ATA (hard disks, Compact Flash)/ATAPI (CD, DVD) for the VIA686b chip which are on the mother charts of A1SE, A1XE and µA1. Then when the problem of interaction was included/understood, I made a pilot for chip IDE Silicon Image 0680, which offered to the users of A1XE a solution to be able to use DMA IDE at the same time as the Ethernet chip. Then when a solution was found for the problem of interaction, I made evolve my pilots to Serial ATA. Today, all chips SATA Silicon Image NCV are supported by my pilots, the 3112, the 3512 and the 3114.
I also produced some other tools (like "idetool" allowing to regulate all the parameters of these pilots) and bookshops (like a wrapper OpenPCI allowing to use the pilot of the chip Ethernet RTL8139 of Benjamin Vernoux under OS4, or the software OpenPCI de Guido Mersmann, initially written for Amithlon, like AmithlonTV etc).
I carried my own applications (AmiVNC, swrender Warp3D and Skulpt) under OS4.
I form also part of the "under-team" Warp3D/Nova, and I am satisfied with the course taken by the things on this axis. I have also some other projects OS4 on fire, but I will not say any more...
# could you tell us what you know of the genesis of the OS 4.0? Why Hyperion did it take this route, whereas there was already work completed on MorphOS?
No comment there above:
Like known as above, I was neither active nor spectator of the initial genesis of OS4, before integrating the team.
I have any point of view on the MorphOS alternative, being never interested it having never used, and me of close with his origin, his design or his functionalities.
It almost never sometimes happened to me to give an opinion on MorphOS, neither in public nor in deprived, except for some exceeded remarks when the flood of inanities which attacks any developer or user OS4, overflows.
In any event, I do not know anything in MorphOS, I cannot thus have legitimate point of view (although there are some who do not stop with that). As for his history, I consider that it was held apart from the Amiga world, it thus does not interest me, as well as BeOS (never considering), AROS (never considering) etc. I have already enough OS4 to occupy itself.
# Which were the relations with Amiga Inc and the other teams which work on parallel projects (MorphOS, AROS...)?
Amiga Inc never intervened not in the development of OS4. Amiga Inc conceded in Hyperion a licence of development of OS4 for the machines Classic/PPC and AmigaOne. In this perimeter, one "does what one wants", while endeavouring to do it the best possible one.
Amiga Inc is focused on its own AmigaAnywhere technology. I do not know if OS4 will play an unspecified part in this strategy in the future.
I am not in the details of the relation Amiga Inc/Hyperion, but I do not feel any concern for the continuation of the events. I suppose that one will be able to continue to develop OS4, and that it will be necessary to negotiate with Amiga Inc if one wish to carry it on other platforms that those for which Hyperion has a licence.
I do not know any (nor does not think about it) not more.
As for the other "parallel" systems to take again your words, I regard them indeed as "parallels", i.e., as would say my mother, than they will never cross our trajectory. I do not carry any interest to them, simply because, as I mentioned above, OS4 occupies me sufficiently (if not, why also not be interested in Syllabe, ReactOS, etc?).
# Which is your opinion on the division of the community in two blocks?
I do not include/understand why these two blocks interfere, since their trajectory is "parallel". What surprises me more, it is omnipresent aggressiveness towards OS4, of those which did not choose this way. Why are they interested in it? Personally I do not take any interest in MorphOS, and we never evoked it within team OS4 (with share into private when I made MOSemu with another of the developers, the motivation of then being reduced "to prove the concept").
In fact, I think that there are in the "autre block" the usual components of a community Internet, in particular some fanatics who make the essence of the noise, with for some a behavior of irreducible aggressive agitation, that they exhiberaient in any event in any context. Are the remarks of the kind "the existence of OS4 a crime", or even "AmigaOne does not go" which flood any discussion on any subject having milked with OS4, or other insults which are addressed to us are more or less easy to support according to mood, but finally they say some longer on the side neurotic (worrying?) their authors that on anything else.
# Comment on the how the OS 4.0 team is organized, currently?
A project leader: Ben Hermans. A technical person in charge: Hans-jörg Frieden (itself developer). And 42 other developers.
# How many are paid and how many voluntary members?
Only both Frieden are "employees". Others have a contract at the end which they touch some rights, but a broad part is voluntary, like me.
That has a principal advantage and a disadvantage:
Favour: one is motivated only by "making well".
Disadvantage: it is difficult to manage a project of this type, particularly under the angle of the "deadlines":).
The word of Frieden summarizes this state of mind well: "When it's done".
# Which are the tasks in progress?
To date, mainly of the finalization. In fact, there remains only one single module which was not remakes out of C and properly carried PPC: console.device, and it is in the course of finalization.
It remains much of things to be made so that we considered that all is finished, but that will leave us work and motivations for OS4.x:).
# Which are the strong points of the OS 4.0 compared to AmigaOS Classic and compared to the great systems of the market?
The strong points are in my opinion:
Reactivity. I have recently MacMini, the most powerful model available, and there are many situations where one waits... Even a little, but especially without feedback (kind, the pointer of waiting starts to turn only at the end of some tenths of a second).
The simplicity of administration. One has textual files or XML for régire 99% of the operation of the system, with a little knowledge, one can analyze a dysfunction, and to rectify it, whereas once Windows XP and its base of registers, or Linux and its gnome/kde tools all-go-well-such an amount of-that-all-go-well leave in gimlet, even with much of knowledge one is likely all to finish in reinstalment...
On the other hand, there are as large weak points, as it is the delay in terms of fundamental concepts (protection memory, streaming, firewire, etc), or of applications (browser Web, office automation continuation, Java etc).
Everyone knows it, but that does not remove with OS4 its interest in terms of hobbystic platform alternate (nor even in terms of other more professional applications which could exist, on which I do not have any point of view).
# What could motivate a former amigaïst, or a user of PC/Mac, to cross the step?
To date, probably interest hobbyst of a different approach, as well as other OS alternate.
# Which are the large projects to come on OS 4.x (software and evolution of the system)?
I do not know... We have a mountain of wishes technical and functional, they will come "when its done", I cannot really say some more.
But one can nevertheless note:
-That some bricks have gild and already posed to facilitate the bearing of applications OpenOffice kind.
-That we have almost all the tools necessary for builder the applications Posix (of which a clean reimplementation of ixemul/ixnet, a native Perl, a native Python, etc).
-That formal specifications exist gild and already for certain bricks to come, like Warp3D/Nova, which gives up the antiquated concepts of current Warp3D to embrace the functions of the programmable GPU (famous "the shaders" necessary to the bearing of all the recent plays, even with the writing of original applications 3d drawing left the chips post-voodoo3), or a system of management of the memory which will replace the current system.
-Last but not least, that I do not see any reason absolutely so that the team does not continue her work beyond the release 4.0.
# And now, defects of the OS 4.0 and AmigaOne. Initially, could you give a progress report on the technical problems of the charts?
Yes. Here the inventory of fixtures:
1. (rare) A1 G3SE (those which has a PPC 750cxe with 600 MHz welded onto the mother chart) have problems of various natures, they are really to regard as partially functional prototypes having preceded XE.
2. A1 XE have the following problems:
2.1 The AC97link connection between the audio VIA686b and chip Sigmatel does not function. Functional consequence: no audio without chart NCV. Solution: problem which cannot be corrected.
2.2 The chip Ethernet 3Com and VIA686b intéragissent at the time of transactions IDE UDMA. Functional consequence: freezing of the probable machine at the time of a simultaneous transfer IDE DMA with an Ethernet transfer. Solutions: problem which can be corrected, two tracks to be crossed, two wire to be welded; if not, one can use one of supported charts NCV IDE (ATA or SATA), or even use it VIA only in mode PIO (much slower).
2.3 Resistances of 15 KOhm were forgotten in the wiring of ports USB. Functional consequence: the USB functions, but does not detect insertions of hot peripherals. Solution: addition of 8 missing resistances.
2.4 The Articia chip does not ensure indication PowerPC when a transfer DMA on the bus requires to invalidate the mask of PowerPC ("mask snooping", "hardware hides coherency"). Functional consequence: none in OS4, since under AmigaOS, it is responsibility for the pilots to ensure this coherence of mask. This point does not affect any the performances, with the passage, because the instructions of emptying of the hiding place used by the pilots before starting a transfer DMA have the same result as the emptying caused by the snooping in the medium of a transfer, only the release is different (software in a case, hardware in the other); those which are really interested to know some more can contact me. On the other hand, for Linux, it is advisable to use adapted pilots and/or a kernel.
3. (very very rare) µA1 MK2 has the same problems as XE with in more slot NCV not completely fonctionel. Same consequences, same solutions. I doubt in any event that there is in France, in so far as I remember it, only 7 were manufactured, they must be regarded as partially functional prototypes having preceded the µA1 "MK3" or "C" which are those that everyone connait.
4. The µA1 MK3 or "C" (it is those which are on sale currently) have only the problem of coherence of mask, which does not have any consequence under OS4.
# Concrétement, on the level of the end-user, how does that appear, as private individuals with the new charts µAOne?
Under OS4: no demonstration.
Under Linux: need for using a kernel and/or pilots who ensure the coherence of mask in a material way before starting a transfer DMA. But I think that that which wants to use Linux finds in any event much better counts in a machine x86, less expensive, supported better than any PowerPC machine.
# a very polemical subject appeared these last weeks, in the forums, about protection memory and of the resource alignment of the OS 4.0. Coarsely, of what does it act? How the things will evolve/move in the future?
It creates for itself polemic tous.les.jours about AmigaOne and OS4... apart from the AmigaOne community/OS4.
With regard to protection memory and the resource alignment, I gave my point of view in an answer to Jerome Senay in one of innumerable the threads AmigaImpact to polemical vocation supplied with the band of usual fanatics. I thus quote myself below:
1. In OS4, as announced and without more result than the initial ambition, there is a degree of protection memory. Inter alia, it is interdict to write in a segment of code. We are in an intermediate stage, for which a total layout MMU was introduced into the system (AmigaOS, hein, not a vaporous trick located at the lower part, or the top, or on the side, nobody know too well, of which nobody will never be useful himself), aiming at satisfying our objectives on this level, which is mainly to add a swapper, able to discharge on disc a memory page not recently used, to use it elsewhere and screw poured. This swapper is functional. It will perhaps not be diffused like such (still that, one will see) because we are in addition completely remaking the system of management of the memory of OS4, to go towards the following stage, which will be a layout MMU by task, and at this time there one will speak again about it.
2. In os4, as announced, there is an amount of resource alignment. We added a certain number of new functions to the various subsystems, aiming at encapsulating the old internal functions of allowance to these subsystems, by adding to the passage the entry of the resource allocated in a list suitable for the task which allocates, in order to be able to release the resources not released by the task itself with its suppression.
It is easy to check, the documentation of exec.library (for example), described a function "AllocSysObject" who allows to allocate (and pre-to initialize, with the passage) in a trackée way. A rapid examination of includes/exec/exectags.h will make it possible to see the list of the items which are supported by this method to date:
* IORequest *
* Hook *
* Interrupt *
* List *
* DMAEntry array *
* List node *
* Message port *
* Message *
* Signal semaphore *
* TagItem array *
* Memory Pool *
* Item Pool *
Naturally, all that is written by developers OS4 endeavours to use these functions, and benefit draws some.
But that will not make the old 68k applications more intelligent.
# A present, and in a purely speculative way, your forecast of exit of the OS 4.0?
The date is stopped since beautiful lurette, but I am not entitled to you to communicate it. I think that it will take all its direction in effective times;).
# Speculation still: you think that OS 4.x will find how many users, in the long term?
No idea, let us say a few thousands.
# Which is the commercial viability of the OS 4.0 and the perenniality of AmigaOne?
Concerning OS4, I would not have any legitimacy with gloser on the subject of his commercial viability, not having never seen the plane business of Hyperion nor that of Amiga Inc One can simply note factuellement that the dead point making it possible to balance the accounts is not that of Microsoft (report/ratio to the number of paid developers), and that the ambition of the team is not either that of Microsoft.
# the survival of the OS 4.0 is completely related to Hyperion... as long as Hyperion exists;) do you know which is currently the activity of the company?
Not, no idea. I suppose that it would be necessary to raise this question with Evert Carton or Ben Hermans.
# What would occur if Hyperion disappeared?
Eh well, OS4 would certainly return in the bosom of Amiga Inc AmigaOS is a credit of Amiga Inc, it is a fact which seems badly included/understood, but which explains current operation perfectly: Amiga Inc concedes rights on this credit, within the framework of contracts.
# Speculation always: what does one have to again await from Eyetech in term materiel?
Absolutely no idea, I suppose that it would be necessary to raise this question with Alan Redhouse.
# Garry Hare, the chairman of AInc, with evoked a material, which is not AmigaOne, but which made turn the OS 4.0... Your opinion on the material in question?
Absolutely no idea, I suppose that it would be necessary to raise the question with Garry Hare. I imagine that it probably acts of a "tablet PC" containing PowerPC, or of a set signal box... no idea, really.
# What do you think of Amiga Inc?
I mentioned it above, these people have a strategy centered on the development of their technology Amiga Anywhere, which in the current state of the things, does not have any relationship with OS4. Amiga Inc never interfered in the development of OS4.
Their bond with OS4 is as follows: they are owners of the rights relating to AmigaOS, therefore nobody can make AmigaOS without their downstream. Hyperion obtained this downstream to make OS4, for the machines Classic/PPC and AmigaOne. I do not know the contents of the contract, with beyond that.
I do not know at all what it will occur then.
# did you test MorphOS on Pegasos?
Not. I saw MorphOS only one time on Pegasos during 5 minutes, when my friend Benjamin Yoris brought his Pegasos to the office. I retain this "interview" that the equivalent of Workbench was skinné, but not very functional. One can also decorate OS4, but I am not at all faded look gnome or MacOS. Personally I like the personality of the look owing to lack of OS4, but it is typically a ground of personal sensitivity.
Once again, I regard MorphOS as another OS, which has nothing to do with the OS which interests me, as well as BeOS etc; I do not have simply enough spare time to interest me in another OS.
# Towards what do you think that MorphOS goes evoluer?
# What does make you remain in communautée Amiga in 2005?
The passion of a hobby interesting and the participation in a project carried out by a team which works since 2 or 3 years without never to have crossed conflict, a team of gentlemen.
note well: OS4 is only a hobby among others, I have the same kind of relationships to the team which develops PS2sdk, or the still active community of developers APPLE 2, without counting all that interests me but which has nothing to do with data processing.
# do you have a question which I forgot to ask you and auquelle you would like to answer?
Euh... I spoke already much:).
# a last message for our readers?
I am been willing to converse with all those which carry a sincere interest with OS4, that it is as a user, or with the title of curiosity. That they benefit from it, there are only two French in the 43 quoted developers!
More information on some of the personal projects than I quoted in this text are on a Web site than I put unfortunately up to date only from time to time, and who am very incomplete: s.guillard.free.fr.
Last edited by cecilia on 22-Apr-2005 at 11:43 PM.
"In terms of worship, I worship the God of Irony.
That's the only God that I know exists." Terry Gilliam