News and commentary from the cross-platform scripting community.
Mail Starting 4/4/97
At 9:16 PM -0800 4/5/97, John Gilmore wrote:
From: firstname.lastname@example.org (Chuck Shotton);
Sent at 4/6/97; 2:27:03 AM;
>That was a nice piece of FUD, Dave, but what is the point?
But not nearly as blatant as the FUD Sun threw out all last week. I've never seen so many scared people on stage at one time. And I really wonder if they actually think they're doing their supporters any favors by scaring the hell out of them with the Microsoft Boogey Man. Their rock-throwing at Microsoft was childish at best and an outright disservice to the conference attendees at worst.
>The one thing Sun has insisted on is that all variants of Java stay
Hmmm. Listen to yourself, John. "The one thing *Sun* has insisted on..." So let's see. Is Java an open standard, freely available, or is it something Sun owns? If it's the former, then Sun has no place insisting on implementations conforming to *its* ideals, much less setting itself up as its own standards body. If it's the latter, then the tiny monoculture of Sun employees thrashing out bits of Java and APIs can't possibly understand or accomodate the wide range of needs and requirements of all Java developers. It's typical NIH arrogance that they're guilty of, not some altruistic desire to do what's right for the language. Since it appears that the latter is actually true, that Sun owns Java in spite of what they say, I have to say that the fate of Sun and their language seem doomed to follow in the footsteps of other great NIH FUD-spewing companies like Apple. Sure sounded like Apple employees on stage last week at JavaOne and comments from strangers in the audience around me echoed my sentiments.
>I advocated strongly that Java be free software (in the GNU sense).
>That anyone could build innovations on top of it, if they released
>the source code for them.
I agree 100%. This is PRECISELY how Java should be made available. Making the source to C compilers, Perl, TCL, various Unix shells, etc. has not resulted in an unmanagable set of variations for any of those languages. It seems alarmist for Sun to claim it would be otherwise with Java. We're all big boys and girls. We can organize a standard just as well as Sun can, and it'd be a lot nicer if you didn't have to plunk down a big wad of cash for a Java license just to get your e-mail answered when you send in language and API changes.
>Sun decided that they needed stronger legal
>connective tissue to reach their primary goal of "write once,
>runs everywhere". They were probably right.
With this I totally disagree. Sun is free to pursue their OWN implementation of Java. If one of their requirements is "write once, run anywhere", then they might end up with a tool set that appeals to people who do a lot of cross platform development. But *I* (or Microsoft or Borland or Symantec or Metrowerks or anyone else) shouldn't be precluded from implementations that vary according to our own requirements. I say let the market decide what's best, not one company with some input from a few of its licensees.
The only other language I have seen receive such tight controls and limited access to changes and the reference implementation was Ada. Look at what that got that community...
That was a nice piece of FUD, Dave, but what is the point? The one thing Sun has insisted on is that all variants of Java stay the same. I.e. that there be no variants. Of course Microsoft would love to have a scripting language that had the market presence of Java but only ran on Windows machines. Ditto DEC for Alphas. Ditto IBM for OS/2. Everybody wants to lock their customers in. Everybody wants to think they "own" Java. Sun has this tendency too, but at least the guys at the top DO understand the value of open systems and how supporting that concept has raised their company from a tilt-up shed to the big leagues.
From: email@example.com (John Gilmore);
Sent at 4/5/97; 9:16:58 PM;
I advocated strongly that Java be free software (in the GNU sense). That anyone could build innovations on top of it, if they released the source code for them. Sun decided that they needed stronger legal connective tissue to reach their primary goal of "write once, runs everywhere". They were probably right.
Your piece on Ellison's lukewarm appraisal of Rhapsody and preference for System 7 is very interesting. I haven't seen the press reports you referred to but I can tell you Ellison's head is in the right place, even though it might be for different reasons than I believe.
From: firstname.lastname@example.org (Paul Mercer);
Sent at 4/5/97; 5:27:47 PM;
This isn't well known in the usual Mac circles but there really is a technically feasible way to move the Mac OS forward, to keep it alive and maybe even grow it a bit. The approach is called Virtual Mac and I invented it in the summer of 1992. I tried continually to pitch it to Apple till I left in 1994. For a while in 1993, when I reported to Dave Nagel, it was even a funded project. But in the end, Apple eschewed Virtual Mac in favor of a grandiloquent play called Copland (and now Rhapsody.) Apple loves big projects: Pink, GX, AppleMail, OpenDoc, Kaleida...
To put it succinctly, Virtual Mac is a technique in which new core OS functionality, including preemptive multitasking and separate protected memory spaces, can be added to System 7 while maintaining a high degree of compatibility with third-party and internal Apple system software. Obviously, external compatibility important but it turns out much of Apple's internal software is technically indistinguishable from the third-party stuff. Apple will be finding this out as they move forward with Rhapsody.
Most of the old timers in the Mac system software community is aware of Virtual Mac and pretty sick of seeing me try to rally support behind it. At least no one has been able to shoot it down architecturally. It can even be done outside of Apple but no one has the courage. And who can blame them given the sorry state Apple has been in the last couple of years.
On a note tangential to Ellison's Mac OS thoughts, in 1991-1993, I felt strongly that the Mac OS could be a viable platform in which to deploy consumer electronics products and we had 2 projects (Rolex & Swatch) that explored it. Again, Apple (Sculley that time) chose the grander play, going with Newton and PIE. Now Apple is fighting for its survival and no longer concerned with such adventures. That's too bad since Mac technology in the early nineties had some real design wins. Today, Moore's law has caught up with the Microsoft and they can build and sell Windows CE machines for $500.
Enough history. Onward with Java!
I see the Java situation just a bit differently.
From: email@example.com (Larry Tesler);
Sent at 4/5/97; 3:00:10 PM;
Re:"Hail This! and Bill Gates on Java"
Sun believes in defining an open API and letting companies compete on implementation. If Microsoft has a better-in-some-way implementation of the Java VM than does Sun, Sun has no objection. Sun just works harder to improve their own VM, which may require them to buy a company with helpful technology. The same goes for Apple and other VM implementors.
The serious battle is not about VM's. It is about Java application access to native API's. In applets, Sun wants no native API access. Every applet should be secure and should behave identically everywhere. In Java applications, Sun wants the usage of native API's to asymptotically approach zero over time, as the Java API's are fleshed out. Microsoft wants the usage of API's named "Windows" to asymptotically approach 100% over time. So as Java becomes popular, the conflict between Sun's and Microsoft's goals will grow.
But I think all OS vendors agree that some apps will be platform-specific and highly tuned to optimize personal productivity, while other apps will be platform-neutral to minimize the porting issue for developers, systems managers, and users. Agreement about that dichotomy is the common ground among us all.
I also think nobody disagrees that there will always be lots of languages in use. Object models will become increasingly compatible so that different languages can be used together seamlessly.
I have been a partisan of OODLs (object-oriented dynamic languages) since Alan Kay showed me Smalltalk in 1972. OODLs I have known, loved, and worked on include Smalltalk-72 through 80, HyperTalk (a semi-OODL), AppleScript, and Dylan. I'm also pretty familiar with ScriptX, SK8, and Objective C. But when I first saw Oak around 1993, I liked it better than any of them. And when Java hit the front pages in 1995, it took me only a week to convince most of my colleagues that these other OODLs were going to be overshadowed.
Sure, other OODLs can run outside web browsers. But how many people will know them better than they'll know Java? Java is the OODL that will be taught in most high schools and colleges and that will be learned by most self-taught web-page programmers.
I agree that if you have a great software idea, and if Java isn't yet up to the task, you should use something else. But I don't agree that Java is a distraction from the web. They are orthogonal, with a compelling intersection.
PBS News Hour, arguably the best news program on television, ran a story yesterday on the state of the economy, amid rosy reports of how vibrant it is. Among the guests from representative industries was the head of Sun (I forget his name, isn't that inte resting. . . ). He mainly commented on the overall growth in the technologies sector (two decades of growth, at least).
From: firstname.lastname@example.org (Matthew Sinclair-Day);
Sent at 4/5/97; 12:20:12 PM;
Sun, Java et al
But the interesting comment he made was that the industry is going through a consolidation phase right now and might emerge with three or so Big Computer Manufacturers. He likened the whole thing to the consolidation of the automobile industry way back wh en into the Big Three. Well, of course that makes one wonder if he and Apple are talking, but I'd like to comment on the consolidation itself.
He said, software writers will flourish because the technology is evolving to allow people to write without being tied to a "publisher". We're already seeing this in the Frontier world and elsewhere. His use of "publisher" and "writer" are interesting ter ms, as they broaden--or in today's parlance, "upgrade"---their meanings to account for new technologies. Someday, I suppose, the convergence between "software" and "writing" will be complete. I'm not sure if we're going to want that. . . . but we'll see.
On the hardware side the economics will require consolidation, he said. I'm in no position to evaluate this statement, but I can say that the analogy with the Big Three auto makers does not really assure me that the future will be much different from wha t it has been. Big Money, Big Players, Posturing, and Engineering not necessarily for the sake of art or good engineering. . . . yuck.
About your comments on Java, I just don't understand the frenetic energy in the "market" over the language, over something which promises much but is only now beginning to deliver. What faith people have! I wonder, is this faith essentially a "secular" fo rm of the faith people have in achieving deliverance during the upcoming millennium?
You're comments on Java were heartening. It's the idea which matters. I'm not much of a programmer per se, but I like scripting Frontier and dabbling with Smalltalk Agents. Frontier and Smalltalk allow me to create what I need and want quite nicely, tha nk you.
Here is another interesting Java tidbit. Sun seems to think that they invented Write Once Run Anywhere. Well, I would claim that K&R did this many years ago with C. You can get a C compiler for just about _any_ machine, and the first C program that I wrote ("Hello World"), or for that matter that most anyone wrote will run on all of them. Stick to standard C I/O and the stdlib and you really can write once run anywhere. In fact, stick to ANSI C and I bet you can support an order of magnitude more different types of machines than Java can right now.
From: email@example.com (Alex Hopmann);
Sent at 4/5/97; 1:47:21 PM;
Write Once, Run Anywhere
Of course this leads to Bill Gates point- Who wants to write software nowadays using only the standard C libraries? Or more importantly, who would want to use that software?
It would seem from your comments that you were present for Eric Schmidt's afternoon keynote featuring, first, Vint Cerf, Bob Kahn, et al representing the past of the Internet, and Gosling and Joy representing the present with Java. When asked about the future of the Internet, Kahn responded that he saw the future of the Web in content that was object-oriented, structured, and dynamic.
From: firstname.lastname@example.org (Steve Holtzman);
Sent at 4/5/97; 12:42:51 PM;
I would like to (re)call your attention to the demonstration of Perspecta's radical new interface that Miko, the Java evangelist, made as part of the session representing, in his words, "The future of the Web". A Java-client displayed information organized in a 3D information space. Miko simply moved toward a topic of interest and sub-topics and ultimately articles appeared delivered in real-time across the net. As he moved toward one of the articles, it launched the web-browser so the content could be viewed. He then returned to the information space to pursue another topic of interest, organized from a completely different perspective.
While not trying to be entirely self-serving, I think you need to think out-of-the-box. You mention Netscape, Sun, IBM, Microsoft and Apple. "Microsoft has a deep base of know-how in easy to use graphic interfaces," but it is defined in terms of the desktop. Apple and Capps, similarly, have designed systems that were built with a different world from the Web as their design center. The Web, however, doesn't fit neatly into a desktop metaphor:
The future of the Web, and Java, will not look like Windows or the Mac.
And the future of the Web -- to the extent it requires entirely new paradigms -- may not come from these desktop-tradition-anchored sources. While at Perspecta we believe we are presenting an entirely new interface for approaching the organization and navigation of information on the Web, I'm sure other young start-ups are also working on other fresh approaches. The future of the Web will be defined in terms of the desktop-anchored past.
President & CEO
You're right about Yogi, Dave. Another great quote: "No one goes there anymore; it's too crowded." Thanks for the insights, keep 'em coming. -- KC
From: email@example.com (Kevin F Compton);
Sent at 4/4/97; 6:49:37 PM;
Remember the Intel Inside campaign that made people care about what processor they had?
From: firstname.lastname@example.org (Jim Louderback);
Sent at 4/4/97; 12:00:00 AM;
Well I predict we'll get a "Java Inside" campaign that will attempt to make users, consumers, etc care about what language their application is written in. Will it work? I think not, but it will be fun to watch.
I think I wrote this a month or two ago: "Java's not a programming language, it's a religion".
(I'm not anti-Java, but it really is just a programming language attempting to become a platform -- and your toolbook piece was right on)
Look at the ZDNN coverage of Java. We spent time with Bill yesterday. They are schizoid on the issue (http://www5.zdnet.com/zdnn/content/zdnn/0403/zdnn0005.html)
Perhaps interesting to note: the same people engaged in the Rhapsody "Hail Mary" are the ones behind the Netscape/Sun "Hail Mary." IFC was created by ex-NeXT engineers and bears a remarkable resemblance to NEXTSTEP and the AppKit.
Sent at 4/4/97; 5:17:48 PM;
Common link between the "Hail Marys"
For the record, I disagree with you about both "Hail Mary" plays.
Apple needs a new OS to carry it forward. The market has spoken, you're being willfully blind if you deny this fact. No one - not Microsoft, not Be, not IBM, not Sun - is promoting an OS as outdated in its fundamental system services, multitasking, memory protection, multithreading as is the MacOS.
And the last thing that Sun and the Java world needs is the chaotic world of hundreds of competing "standard" UIs - X Windows "mechanism, not policy" all over again. This is just one of the reasons that the Unix world is being slapped silly by Windows NT in the low end where the two meet and compete.
If Larry Ellison feels that Rhapsody is a loser, it's probably coming from him as a solid Mac owner and user; those are the most confused and disheartened people these days. But, hold on, there might be more understandable and complete strategy communications soon that will give Mac addicts, who have been true for so long, a ray of light and a reason to want, instead of fear, Rhapsody. That something inside can eventually sway large numbers of Windows users over. Isn't that one of the things that makes Java a pleasant idea? Please be patient with Apple for a little longer.
From: email@example.com (Steve Wozniak);
Sent at 4/4/97; 11:01:11 AM;
I don't know Apple's stated position but our attitude toward the Mac should be "Macintosh Forever", meaning that as long as our users want more (HW and SW) they'll get it. I'm very sorry about the message conveyed by dropping work on certain technologies recently, but in these troubled times it might be better to err on the side of cutting back a little too much.We'll also work full bore on bringing the nExt stuff to the Mac ASAP.
Two more sayings from Yogi:
From: firstname.lastname@example.org (Daniel Berlinger);
Sent at 4/4/97; 12:57:32 PM;
The future ain't what it used to be...
When you come to fork in the road, take it!