20-Apr-2024 09:13 GMT.		 UNDER CONSTRUCTION
Anonymous, there are 39 items in your selection
[Forum] A big step forward in cross-platform computingANN.lu
Posted on 15-Sep-2004 22:32 GMT by Gary Goldberg39 comments
View flat
View list
By Leander Kahney 02:00 AM Sep. 13, 2004 PT A Silicon Valley startup claims to have cracked one of most elusive goals of the software industry: a near-universal emulator that allows software developed for one platform to run on any other, with almost no performance hit. Transitive Corp. of Los Gatos, California, claims its QuickTransit software allows applications to run "transparently" on multiple hardware platforms, including Macs, PCs, and numerous servers and mainframes... http://www.wired.com/news/technology/0,1282,64914,00.html?tw=wn_6techhead
A big step forward in cross-platform computing : Comment 1 of 39ANN.lu
Posted by Lando on 15-Sep-2004 23:18 GMT
"almost no performance hit"

I don't believe that for a second.

"QuickTransit fully supports accelerated 3-D graphics and about 80 percent computational performance on the main processor."

Right, so I can run my OS X apps on a 3.6Ghz x86 box at the speed of a 2.88Ghz G4 or G5?

I think this is just marketing hype and if it lives up to their claims I'll be very very surprised.
A big step forward in cross-platform computing : Comment 2 of 39ANN.lu
Posted by Andrew Korn on 15-Sep-2004 23:27 GMT
Isn't Transitive the company that were touting a dymanic translation system under the name "Dynamite" a few years back? It was discussed somewhat around these parts when Intent / AmigaDE were more of a hot topic.
A big step forward in cross-platform computing : Comment 3 of 39ANN.lu
Posted by Joe "Floid" Kanowitz on 16-Sep-2004 01:11 GMT
In reply to Comment 2 (Andrew Korn):
Isn't Transitive the company that were touting a dymanic translation system under the name "Dynamite" a few years back? It was discussed somewhat around these parts when Intent / AmigaDE were more of a hot topic.

Perhaps you're thinking of the Dynamo research project from HP?

Anyhow, I can believe they peak around '80%,' whatever 80% exactly means (lower numbers are, of course, more plausible). This isn't rocket science, it's just the equivalent of WINE combined with a CPU emulator/translator someone's bothered to optimize. Why wasn't it done before? Probably because it's much less of a pain in the a** to go out and buy hardware.

Meanwhile, it's not like this is a perfect solution or anything. UNIX-to-UNIX is relatively easy (hey, comes free in every BSD), the novelty being, again, whatever this miracle bytecode translator they claim to have... On the other hand, people have been struggling to reimplement the Windows API for years (WINE, PE2LX/Odin), and you'll notice that while they can *demo* it, they sure aren't selling it just yet. OS X poses a similar challenge, as comes to dragging over the whole shebang of Quartz.

Figure this is mostly a way for corporates to migrate off 'dead' platforms (PA-RISC, Alpha, some flavors of #?Sparc... PDP-n?) and less-than-kicking OSes (OSF?) to the de-facto AMD64/[Linux|BSD] standard, while reaping some performance over crappier emulators (assuming reasonably non-hairy code). And that, given some revenue from that, they *should* be able to spit out Windows and Mac support within a reasonable timeframe, but the compatibility lists will be somewhat limited and the modules will be maintainership hell.
A big step forward in cross-platform computing : Comment 4 of 39ANN.lu
Posted by Joe "Floid" Kanowitz on 16-Sep-2004 01:18 GMT
In reply to Comment 1 (Lando):
"QuickTransit fully supports accelerated 3-D graphics and about 80 percent computational performance on the main processor."

If you consider that most 3D software on *NIX will be built targeting GLX and X11, the former almost becomes almost easy.

Right, so I can run my OS X apps on a 3.6Ghz x86 box at the speed of a 2.88Ghz G4 or G5?

...Given that line of thought, my suspicion would be "Yes, but probably only if they're 'Darwin' apps targeting X instead of Quartz."
A big step forward in cross-platform computing : Comment 5 of 39ANN.lu
Posted by Anonymous on 16-Sep-2004 01:46 GMT
Completely offtopic, not amiga news...
A big step forward in cross-platform computing : Comment 6 of 39ANN.lu
Posted by AMC258 on 16-Sep-2004 03:17 GMT
"Typically with emulation you take a big performance hit," he said. "Their big breakthrough is they are much more efficient ... but there's so much overhead anyway, you can pretty much put any software on any platform. The power user might notice the difference, but the other 95 percent won't notice."

This puts Amiga into context. There is so little overhead in the Amiga operating system (I can't speak for OS4 though) that Amiga apps are likely to run much slower than their quoted 80% on a "per-MHz" basis.
A big step forward in cross-platform computing : Comment 7 of 39ANN.lu
Posted by Chris Perver on 16-Sep-2004 05:50 GMT
Isn't that what Amiga was trying to do?

Oh that's right, they started developing software for phones. Whoops
A big step forward in cross-platform computing : Comment 8 of 39ANN.lu
Posted by Anonymous on 16-Sep-2004 06:16 GMT
Yeah gotta love Quake3 on a 3270 display. Near Native! Runs On Mainframes!
A big step forward in cross-platform computing : Comment 9 of 39ANN.lu
Posted by bennymee on 16-Sep-2004 06:28 GMT
In reply to Comment 1 (Lando):
>Right, so I can run my OS X apps on a 3.6Ghz x86 box at the speed of a 2.88Ghz G4 or G5?

I don't think so, because you can't compare MHz. And don't forget the IPC.
A big step forward in cross-platform computing : Comment 10 of 39ANN.lu
Posted by Don Cox on 16-Sep-2004 09:38 GMT
In reply to Comment 7 (Chris Perver):
"Isn't that what Amiga was trying to do?"

It's what Tao do, but applications have to be specially compiled to work on Tao's system.
A big step forward in cross-platform computing : Comment 11 of 39ANN.lu
Posted by miksuh on 16-Sep-2004 12:04 GMT
If you forget AmigaInc's errors and problems then this shows that there is nothing wrong in the idea behind AmigaDE. The idea is still good. Software industry is slowly going into direction were you write application once and run it everywhere, it makes sence from the developers point of view. Ofcourse java has done that many years, but Java is not always best solution.
A big step forward in cross-platform computing : Comment 12 of 39ANN.lu
Posted by Don Cox on 16-Sep-2004 13:43 GMT
In reply to Comment 11 (miksuh):
"Software industry is slowly going into direction where you write an application once and run it everywhere, it makes sense from the developer's point of view."

It would if it were true. However, the company making the system has to support every known device, which is not economically possible. For example, Tao's Intent doesn't run on Macs or Amigas.

"Of course java has done that many years, but Java is not always best solution."

Java again does not "run anywhere" but only on a limited number of platforms. Also, it is not an OS, so whether something that requires specific hardware (such as Firewire) works depends on the host OS.

There is nothing new about "run anywhere". ZBASIC and UCSD Pascal attempted it 20 years ago. It can work for a program that has no specific hardware needs, such as a high res screen, a sound card, MIDI, etc.


The other problem is that it does _not_ make sense from the hardware manufacturers' point of view. These guys want exclusive software that runs only on their box. Even the users think like that - Pegasos users would be pleased to have programs that were not available for the AmigaOne, Mac users like to know that they have programs that don't run on Windows.
A big step forward in cross-platform computing : Comment 13 of 39ANN.lu
Posted by Fabio Alemagna on 16-Sep-2004 13:45 GMT
In reply to Comment 11 (miksuh):
The idea was here long before TAO or Amiga even existed, and nowadays the leader of this field is not TAO, it's MS with it's .NET platform (together with its open source implementations) and, to a certain extent, Sun with its Java platform.

Not to mention the Tendra compiler, which emits bytecode as an intermediate language between the machine language and the high level language of choice. Bytecode which could be directly interpreted and/or JIT'ed.
A big step forward in cross-platform computing : Comment 14 of 39ANN.lu
Posted by Anonymous on 16-Sep-2004 14:41 GMT
In reply to Comment 13 (Fabio Alemagna):
I did not say the idea is new. but it does not mean it is an bad idea. You can always try to make things better than competitors, no matter how old your idea is. Ofcourse it is not easy task for the small company, but you can always try.

If I would have to choose between .NET and Java I would definetaely select Java. .NET is still too MicroSoft related.
A big step forward in cross-platform computing : Comment 15 of 39ANN.lu
Posted by miksuh on 16-Sep-2004 14:44 GMT
In reply to Comment 14 (Anonymous):
too much
A big step forward in cross-platform computing : Comment 16 of 39ANN.lu
Posted by sutro on 16-Sep-2004 17:06 GMT
In reply to Comment 13 (Fabio Alemagna):
> The idea was here long before TAO or Amiga even existed, and nowadays the leader of this field is not TAO, it's MS with it's .NET platform (together with its open source implementations) and, to a certain extent, Sun with its Java platform.


LOL

.... ,and nowadays the leader of this field is not Real, it's Micro$oft (together with its open standards WMA format) and, to a certain extent, Apple with its iTunes Music Store and iPod devices.
A big step forward in cross-platform computing : Comment 17 of 39ANN.lu
Posted by Fabio Alemagna on 16-Sep-2004 17:57 GMT
In reply to Comment 16 (sutro):
> LOL

I'm sorry for you, if that's your sole argument. Or perhaps you do have better arguments you care to share with us?
A big step forward in cross-platform computing : Comment 18 of 39ANN.lu
Posted by Fabio Alemagna on 16-Sep-2004 18:00 GMT
In reply to Comment 14 (Anonymous):
> I did not say the idea is new. but it does not mean it is an bad idea.

I don't think anyone said it's a bad idea. I for sure didn't.

> You can always try to make things better than competitors, no matter how old
> your idea is. Ofcourse it is not easy task for the small company, but you can
> always try.

But let's face it: on cell phones there's Java, on desktops there's .NET and Tendra. What market does TAO's technology fit? Who uses it currently? And why? Is it worth its cost?

> If I would have to choose between .NET and Java I would definetaely select
> Java. .NET is still too MicroSoft related.

Parts of .NET, unlike Java, are standardized. Moreover, .NET bytecode and virtual machine is simply better than java.
A big step forward in cross-platform computing : Comment 19 of 39ANN.lu
Posted by sutro on 16-Sep-2004 18:55 GMT
In reply to Comment 17 (Fabio Alemagna):
>I'm sorry for you, if that's your sole argument. Or perhaps you do have better arguments you care to share with us?

Too many to list and I have lots of work to do. Be asured that they are much better than yours. And fortunately, for once, the industry agrees (.NOT with you). Do you expect me to argue with everyone insisting that Newcastle is a bigger/better/more successful franchise than Liverpool ? You are free to choose whatever language to develop. The ground truth is that Java has been here before .NET, it's more "open" then what .NET will ever be and, currently, more popular and better supported in Desktop, Server AND Mobile.
A big step forward in cross-platform computing : Comment 20 of 39ANN.lu
Posted by Fabio Alemagna on 16-Sep-2004 20:41 GMT
In reply to Comment 19 (sutro):
> Too many to list and I have lots of work to do. Be asured that they are much
> better than yours.

If you say so...

> And fortunately, for once, the industry agrees (.NOT with
> you). Do you expect me to argue with everyone insisting that Newcastle is a
> bigger/better/more successful franchise than Liverpool ?

I don't even know of what you're talking about... perhaps it would help if you could stick to the topic at hand.

> You are free to choose whatever language to develop. The ground truth is that
> Java has been here before .NET, it's more "open" then what .NET will ever be
> and, currently, more popular and better supported in Desktop, Server AND
> Mobile.

But you forget that .NET is not a language, .NET is a whole platform based on a virtual machine and a set of libraries aimed at easing development. You can program in whatever language you wish and for which there exists a frontend, with .NET, not so with Java. Yes, both use a virtual machine, but the .NET one is

1) standardized
2) much better at supporting a multitude of languages

Even Java runs on .NET, with comparable or better performances than the "official" Java.

Java may be more widely used, but .NET is simply a better technology.
A big step forward in cross-platform computing : Comment 21 of 39ANN.lu
Posted by Joe "Floid" Kanowitz on 16-Sep-2004 23:39 GMT
In reply to Comment 20 (Fabio Alemagna):
But you forget that .NET is not a language, .NET is a whole platform based on a virtual machine and a set of libraries aimed at easing development. You can program in whatever language you wish and for which there exists a frontend, with .NET, not so with Java. Yes, both use a virtual machine, but the .NET one is

1) standardized
2) much better at supporting a multitude of languages

Even Java runs on .NET, with comparable or better performances than the "official" Java.

Java may be more widely used, but .NET is simply a better technology.

Well, there are always counterexamples.

More pragmatically:

-With Java you get raped by Sun licensing (and to some extents, Sun engineering) up front.

-With .NET, you get all the cheerful flexibility (such as it may be), but with the threat of being torpedoed by MS patents a few years down the line. (Of course, if you're already invested in software, er... monoculture?... for that time frame, it's not a bad deal, considering MS is going to make you ride that pony anyway when they declare the Windows API end-of-lifed.)

It's an interesting problem, to the extent that, if you believe all the bloggish PR, MS never *intends* to bait and switch... Rather, we're supposed to believe the company simply runs in a constant state of panic, and the bait-and-switch/outright illegal business activities only creep in as 'natural extensions' of honest effort -- The Redmond version of the boneheaded mistakes that see other companies focus on A600s or, yes, 'Open' runtimes that aren't open. Which, when you think about it, is a great big advertisement saying "No matter what we say, you can never trust us!"
A big step forward in cross-platform computing : Comment 22 of 39ANN.lu
Posted by JoannaK on 17-Sep-2004 02:40 GMT
In reply to Comment 3 (Joe "Floid" Kanowitz):
It's Dynamite..
http://www.transitive.com/news_10mil.htm

Anyhow.. Anothet day, another Jit compiler... I have no idea if this one is any good or not.. But IMHO it ain't nothing new.
A big step forward in cross-platform computing : Comment 23 of 39ANN.lu
Posted by Johan "Hagge" Krüger-Haglert on 17-Sep-2004 04:35 GMT
"Right, so I can run my OS X apps on a 3.6Ghz x86 box at the speed of a 2.88Ghz G4 or G5?"

No, at the speed of a 2.88ghz x86 running a native version.
A big step forward in cross-platform computing : Comment 24 of 39ANN.lu
Posted by sutro on 17-Sep-2004 13:32 GMT
In reply to Comment 20 (Fabio Alemagna):
Oh My, you do not know what are you talking about do you ? I am not going to spend much time for you, explaining that "Java" and "C#" are languages while "Java" (YES, Sun simply didn't care to come up with a better name for years) and ".NET" are platforms. The Google is there, you can find all the information you need about the why Sun declined several times to "ECMA-standardize" Java even though they considered from 1997 onwards, I am not going to do that for you. The bottomline is that the decision has not hurt the Java platform at all, except in those meaningless flame posts and discussions.

You made some comments that I need not and should not answer. More important issues and differences between the two platforms have been discussed elsewhere. End it here. ANN is for Amiga related flames only (?).
A big step forward in cross-platform computing : Comment 25 of 39ANN.lu
Posted by Fabio Alemagna on 17-Sep-2004 14:06 GMT
In reply to Comment 24 (sutro):
Dude, I'd suggesr you to chill down, and dare put a name on those meaningless words. You like Java, it's ok, I don't. .NET is proving to be a far better technology than Java, for a multitude of reasons that, as you correctly say, this is not the place to discuss.

Whatever is your opinion about Java and .NET, it doesn't change the fact that TAO's technology doesn't show to have any future, as the market for virtual machines is saturated already. Java on servers and cellphones, .NET on clients.
A big step forward in cross-platform computing : Comment 26 of 39ANN.lu
Posted by Kolbjørn Barmen on 18-Sep-2004 11:03 GMT
In reply to Comment 25 (Fabio Alemagna):
I have .NET on my cellephone, as well as Java.
A big step forward in cross-platform computing : Comment 27 of 39ANN.lu
Posted by Fabio Alemagna on 18-Sep-2004 14:31 GMT
In reply to Comment 26 (Kolbjørn Barmen):
> I have .NET on my cellephone, as well as Java.

Of course, cellphones with WinCE will have both :-)
A big step forward in cross-platform computing : Comment 28 of 39ANN.lu
Posted by 3seas on 19-Sep-2004 15:42 GMT
This whole idea of translation regarding computer programs.....

use the source luke....

In other words.... of all the places where translation can be applied then why not do it at the source code level?

I was looking at something recent on linux (gaim???) ... something about ./configure generating the C++ source code needed for my particular system...

I think the only reason not to is for proprietary income generation reasons.

Consider Taos product that requires some translation table/module to allow a new platform to be used with teh same semi-binary.... who profits off that but of course Tao.

Ultimately things will evolve towards maturity in the software industry as the realization that, that which is natural, not patentable, regarding computer programming concepts and datatypes are more direct, simpler to deal with and support innovation.... rather than the fabrication of of some phoney "better way" under some proprietary technology...

In sum total, you can't do better than father physics and mother nature, but only make false claims of such nature.

MS is not an innovation machine, its a marketing machine that take the ideas and works of others and resales it as their property. BASIC and DOS.... where this MS marketing process began...
A big step forward in cross-platform computing : Comment 29 of 39ANN.lu
Posted by noggin on 20-Sep-2004 08:51 GMT
I'm not sure if people actually realise this, but Tao's VM technology is Java. It's a small+fast implementation of the virtual machine, and AFAIK nothing more than that. I'm less clear on AmigaDE, I guess it is either a re-badge with the Amiga name (fair enough), or chucks some game-related API's on top of Tao's Java VM. If they do add API's, they are completely redundant IMO. J2ME (Java 2 Micro Edition) already has perfectly good, standard API's to deal with game stuff already (MIDP, CLDC, MMAPI, JSR-184).

BTW, there are are also more languages that run on the Java VM than there probably ever will on .Net (if your really interested, I'll find you a link to IIRC, hundreds). They are not officially supported by Sun of course, so if you want an officially supported non-Java language VM then just go with MS.

I don't quite understand the point in arguing .Net against Java here, and I certainly don't see .Net as a "better" technology. It's just different. You either you throw your lot in with Sun or with Microsoft. I know what I continually choose based on (1) track record, (2) the relative open-ness of the tech, and (3) being a proven solution. But that is not really the point. The likelihood is that nobody will "win", so there is no point arguing about it until there are independent and fair assessments of the two technologies in *real environments*. This is something we have not seen yet since the only studies so far have been clouded by dodgy back-handers from both sides (some more public than others).

Ultimately, it's too early to tell for .Net - give it a few more years.

For enterprise-level business, and for the mobile phone industry, development using VM-like environments is almost completely Java. That may indeed change over time. It will certainly be interesting to see what pans out.
A big step forward in cross-platform computing : Comment 30 of 39ANN.lu
Posted by 3seas on 20-Sep-2004 10:09 GMT
In reply to Comment 29 (noggin):
Tao has its Java virtual machine but its in competition with other such JVMs..

Taos other product is a Virtual Processor cored product. Create a CPU in software and provide a translation of its instructions to real CPU hardware, where software is written for the virtual processor. This leaves application portability in the hands of those who control the creation, or not, of the translation data from teh VP (virtual processor) to the RP (real processor).

Tao might have more success with their VP if they opened up the translation creation data for others to create. This way a hardware manufacture interested in using Taos VP could create that data themselves (and they would know better what they want too.)

But most important, the Tao VP work is not intended to support a full out desktop OS, but rather smaller embedded systems.

Still though, and in contrast to the VP idea, there are all sorts of flavors of open source linux, even down to small embedded systems and for compiling.... the most used compiler is gcc.... So, again, translation should happen at the code creation level.

Why? What sthe most obvious reason? run time overhead. As a direct compile for the hardware is (no translation during runtime) vs. the translation process during runtime in using a VP.

This VP idea becomes less viable as the direct code generation and compile process improves.
A big step forward in cross-platform computing : Comment 31 of 39ANN.lu
Posted by noggin on 20-Sep-2004 10:32 GMT
Ah, I wasn't aware of their other products. Thanks 3seas.
A big step forward in cross-platform computing : Comment 32 of 39ANN.lu
Posted by Don Cox on 20-Sep-2004 14:10 GMT
In reply to Comment 29 (noggin):
"I'm not sure if people actually realise this, but Tao's VM technology is Java. It's a small+fast implementation of the virtual machine, and AFAIK nothing more than that."

Tao's technology is not Java, and it was invented long before Java. It is basically an intermediate byte-code, as in UCSD Pascal, which is compiled to the native code for each platform when it is loaded from disk to RAM. There are two modules for each platform, one for the CPU and the other for every other feature of the platform.

Java does run on it, and Tao claim a particularly efficient implementation, but so do other languages such as C. Or you can code in Tao's own macro-assembler.
A big step forward in cross-platform computing : Comment 33 of 39ANN.lu
Posted by Don Cox on 20-Sep-2004 14:14 GMT
In reply to Comment 30 (3seas):
"So, again, translation should happen at the code creation level.

Why? What sthe most obvious reason? run time overhead. As a direct compile for the hardware is (no translation during runtime) vs. the translation process during runtime in using a VP."

There is no runtime overhead in Tao's system. The final compile is done as the program loads from disk, before it starts running. It is claimed that the extra loading time is negligible.

It does not translate during run time - it runs native code suitable for the platform.

It's a clever system, just nothing to do with the Amiga.
A big step forward in cross-platform computing : Comment 34 of 39ANN.lu
Posted by Ben Hern on 20-Sep-2004 17:39 GMT
In reply to Comment 32 (Don Cox):
"Tao's technology is not Java, and it was invented long before Java. It is basically an intermediate byte-code, as in UCSD Pascal, which is compiled to the native code for each platform when it is loaded from disk to RAM. There are two modules for each platform, one for the CPU and the other for every other feature of the platform."

Another part of there plaform might be, but their technology is definitely Java too. I think this was what noggin was talking about, i.e. the part that AmigaDE is based on. See their website:

Tao's intent®, Java™ Technology Edition Now a Sun Authorized Virtual Machine

"Over several years intent has been developed by Tao's engineering team for use in connected devices in home and mobile networks. The emphasis has been on creating an open, comprehensive, customizable, compact, fast and binary portable multimedia solution incorporating advanced capability in streaming audio and video. At its heart is Tao's Sun Authorized Virtual Machine and libraries. Today, the software is being deployed in a broad range of markets including digital games, home networking, digital television, automotive and phones."

You also said:

"Java does run on it, and Tao claim a particularly efficient implementation, but so do other languages such as C. Or you can code in Tao's own macro-assembler."

I think you are confusing the other part of "intent", where you can compile the Java language to their proprietary byte code. That is quite different. Their Java VM is an entirely different thing... and it certainly would not get licensed by Sun if it were not a compliant and complete Java solution.
A big step forward in cross-platform computing : Comment 35 of 39ANN.lu
Posted by MarkTime on 20-Sep-2004 18:25 GMT
there has been a lot of misunderstanding about this product on the amiga news groups (no surprise there, since most of you live in candyland).

this is a hardware emulator and some system mapping for unix-like os's....
true, transitive says the don't want to call it an emulator...

but that is purely marketing, it is an emulator of hardware, thats what it does.

Now, as far as WINE...no.

Obviously, if I have a program that calls another program, and that other program doesn't exist...
then transitive can't do anything about that.
So if I have a windows binary, and I call some OS DLL's, but that OS DLL doesn't exist...i.e.
you never installed windows and you don't have a license for it...then that binary WILL NOT WORK.

This isn't an OS emulator at all...if you want to run windows programs, you will have to have a copy of
windows.

Think of the transitive system like another Virtual PC...the technology and speeds may be breakthrough,
but this isn't a software emulator at all, if you need to run windows, you will need to have windows.

Now, Linux is free, and wine is free, and I'm not saying you can't install WINE if you want, but that
has nothing to do with transitive.

Why you all are confused about this...I don't know, because the announcement was clear, and nobody,
but nobody ever said it let you run windows programs on a macintosh without windows.

You can run windows for Intel on PowerPC hardware...but you'll have to have a copy of Windows installed.
If you want to run windows binaries, you will need to license a copy of windows, or run a program like wine.
A big step forward in cross-platform computing : Comment 36 of 39ANN.lu
Posted by noggin on 21-Sep-2004 05:28 GMT
In reply to Comment 34 (Ben Hern):
Thanks for clearing that up for me Ben ;)

I wasn't aware of Tao's other products though. I am generally more interested in AmigaDE (not really on a positive level though) and Java-based solutions in general because that is the area I work in professionally.
A big step forward in cross-platform computing : Comment 37 of 39ANN.lu
Posted by noggin on 21-Sep-2004 05:35 GMT
In reply to Comment 35 (MarkTime):
While that is certainly true, enhancements in technology in this area are very important commercially because of hardware consolidation, not software consolidation. The costs for a moderate server far outweigh the cost for a Windows license and even the cost of the virtualisation software itself (and that is usually pretty damn expensive, but of course you may only need one copy).

If you can virtualise many your servers on one big server, you can really see benefit financially. VMWare, etc. is currently enjoying something of a revolution with big companies wanting to do this.

Anyway, I suspect you knew this, but this is important to highlight too.
A big step forward in cross-platform computing : Comment 38 of 39ANN.lu
Posted by Joe "Floid" Kanowitz on 21-Sep-2004 08:11 GMT
In reply to Comment 34 (Ben Hern):
"Over several years intent has been developed by Tao's engineering team for use in connected devices in home and mobile networks. The emphasis has been on creating an open, comprehensive, customizable, compact, fast and binary portable multimedia solution incorporating advanced capability in streaming audio and video. At its heart is Tao's Sun Authorized Virtual Machine and libraries. Today, the software is being deployed in a broad range of markets including digital games, home networking, digital television, automotive and phones."

Tao's VM runs atop the Elate (well, VP) 'VM,' or you can run javac (IIRC) or whatnot to blow it straight into 'compiled'/translated bytecode. There may have been some various magic in terms of what initiates what, but that's basically it... one interpreter-alike (which may indeed have translated before execution rather than interpreting or JIT-ing, passing Sun's compliance suite is orthogonal to *how* the implementation makes compliance happen), one compiler/translator-alike. Since the VP is a 'middle ground,' the compilation/translation step doesn't take very long, probably gets cached, etc, etc.

I hate to just drop a link to www.funroll-loops.org and run, but I've been through like three drafts of other replies, and they all turn to rant. There's a place for everything, and it's my humble opinion that extremely streamlined instruction-set abstraction probably does belong on many desks. Especially if they'll mostly be used to play Solitaire yet be tied to Windows for it. (Well, rant: It can go either way, you can build your abstraction into an 'environment' unto itself: Elate/Intent and Java to various extents -- or you can build an abstraction 'tailored' to your OS: .NET [though that tailoring is more in the legal realm than anything else], or chances are anything cool anyone does with TenDRA will target either Linux or BSD ABIs, albeit ones that handle the abstracted calling convention*... One approach gets you a write-once-run-anywhere platform that 'becomes' the OS from the developer's perspective, the other makes the existing OS become a write-once-run-anywhere platform.)**

*If that makes sense. In other words, I'm betting anything cool that *would* ever be done with TenDRA wouldn't immediately include the equivalent of the Java standard library or the whole Elate kernel/OS and Intent graphic/media library business... there'd just be the 'translator' itself, which would, er, translate calls transparently. Which, er, is sort of the sane, "monolithic" approach, but demands a portable kernel to begin with (which Linux/BSD are, and XP... remains a clusterf*ck about, to the extent that Microsoft do get an ease with .NET by, one assumes, making the one .NET VM project build-time 'portable' across the similar-API'd Windows Family of Operating Systems... but they've still gone through the trouble to reimplement features 'within' their environment, because they need to eventually use it to replace and subsume everything irreparable in their existing OSes. WTFOMGBBQ)

**So er, yeah, Microsoft are screwing up to the extent that .NET has to be something of an "OS" unto itself because their actual OS ain't no good. When, of course, "just" translating abstract bytecode calling the existing native APIs without creating a layer of here's-some-new-sh*t-for-everyone-to-port-to insulation from them would actually give them an excuse for lock-in/customer retention that doesn't rely on vague dangling legal axes.

So much for this post's coherency. Anyhow, the benefit of all this stuff is *supposed* to be that, even if it costs money, that money buys you freedom, without having to *always* take the tradeoff of "Watching sh*t scroll by for hours." (Hint: Yeah, you already don't have to, in Linux/x86-land. But we're seeing how much fun Flash on Linux/PPC is, right?)
A big step forward in cross-platform computing : Comment 39 of 39ANN.lu
Posted by Joe "Floid" Kanowitz on 21-Sep-2004 08:24 GMT
In reply to Comment 35 (MarkTime):
Obviously, if I have a program that calls another program, and that other program doesn't exist...
then transitive can't do anything about that.
So if I have a windows binary, and I call some OS DLL's, but that OS DLL doesn't exist...i.e.
you never installed windows and you don't have a license for it...then that binary WILL NOT WORK.

This isn't an OS emulator at all...if you want to run windows programs, you will have to have a copy of
windows.

As regards the Transitive product, we don't *know* if they've included anything WINE-like (which takes the place of some core Windows DLLs, enough that you can, so I've heard, run Doom 3 without needing any of Windows installed)* or not. This is because they don't actually *sell* a freaking Windows product yet, as noted.

Presumably their UNIX compatibility modules include *some* essential libraries or functionality to replace them, but that is, indeed, mostly a much easier 'mapping' task. (The fact that it *is* modular suggests that they can do a quick mapping where they can, then reimplement all of Cygwin or EMX and beyond to do the same trick mapping pick-your-*NIX-to-Windows_or_OS/2 if they'd want. Which I'd imagine they'd never bother with, because that's a seriously absurd amount of work for little profit, but if they've been running Windows-on-UNIX demos, they've been playing with the idea whether or not they do need a license of Windows around to support it.)

[So in other words, they can rip out half their product and replace it with something completely different depending on the target. Which is, like I just said, extremely nontrivial, but they're making a point of saying the option's open to do it.]

*If I'm wrong here, I welcome a swift kick. But please confirm it firsthand before you aim it.
Anonymous, there are 39 items in your selection
Back to Top