The Subjectives of OpenDoc
Monday, March 13, 1995 by Dave Winer.
Remember Spiro Agnew? Bias, bias, bias! That was his not-too amusing rant about the press. I expect to hear some of that from the proponents of OpenDoc. But I deny it. I have an OpenMind about OpenDoc. But my mind currently thinks OpenDoc is an open crock, conceived by computer scientists to solve their own problems, to fill their own needs, based on very little true understanding of the world it is being injected into, and it is probably not very open, even if it is very doc.
I am stuck with subjectives because I don't have a copy of OpenDoc to look at, no sample apps to try out, no theories to test. Not that I haven't tried to get a copy. I have. I'm a credible developer. The mere fact that I am not flooded with news about OpenDoc is an important subjective. Sure, they must get the impression that I am unfriendly. That never stopped Microsoft. If you've got Truth on your side, your evangelistic efforts never stop. Apple et al seem to have stopped promoting OpenDoc before they've started.
OK -- next subjective. I know the guys who are doing OpenDoc. We implemented a lot of their specs in Frontier 3.0, because it is a full Open Scripting Architecture (OSA) client and server. The guys who did AppleScript are the same guys doing OpenDoc. What did I think of their APIs? Inefficient, over-engineered, far too elegant, huge gaping holes, and based on incorrect premises about how developers and script writers work.
The "open" part of OSA is a total joke. To be open, they would have to acknowledge our menu sharing protocol, which has become a defacto standard among scriptable apps, and is unique to the Mac, with no counterpart on Windows. Apple, instead, buries its head in the sand, and makes us sell each developer, on an individual basis, on the wisdom of doing the client side of menu sharing. They blow a lot of smoke about standards, but when the dust settles, "standard" means "comes from a large company." [And MacWEEK plays right along, still insisting that developers add AppleScript compatibility. How incredibly cynical.]
I see my own experience and ideas in Microsoft's OLE 2.0, all over the place, after having spent a couple of weeks working with Microsoft people in 1992. Microsoft is willing to listen to developers they respect. They don't seem to care much about the size of the company you work for, or at least their top management doesn't. Apple may be able to listen to developers they respect, but I've yet to meet a developer who has their respect. If you know one, please let me know.
Actually, ironically, I see my own ideas and experiences reflected in Apple's efforts. But these have always been negative appropriations, not win-win deals. Apple steals my ideas, but then uses those ideas to thwart my efforts to get the platform moving in fun, competitive and productive ways. They play the FUD game to the disadvantage of their users and developers. At least when Microsoft plays FUD, they tend to use what already works, instead of making us all wait. [The platform with the most powerful developers is the most powerful platform.]
Anyway, I made a bet with Jerry Michalski, email@example.com, in mid-1994. Jerry will get $100 from me if in 1995, five reviews of OpenDoc parts appear in either MacWEEK or PC WEEK. If four or fewer appear, I get $100. We were trying to come up with a way of measuring the commercial reality of OpenDoc. I think we came up with a good one.
My bet is that OpenDoc will be a clunky API, the subject of much negotiation among the partners, a joyless solution for their problems, not ours. It will have little or nothing to do with the world developers and users live and play in, and therefore won't be well-supported anytime soon by serious (or fun!) commercial development.
I ran into Mike Maples, one of the top guys at Microsoft, at the Phoenix airport last Wednesday. We were both on our way back to San Francisco after attending Esther's conference. I tried out some of my theories about the relative openness and competitiveness of the Mac platform as compared to Windows 95.
It turns out that the 8-character-filename argument works. And so does the story about plug-and-play requiring new hardware. See Positive Energy, 2/3/95. Mike is a very smart guy, and a very biased warrior for Microsoft's cause. I figured that if anyone would be able to rebut these points, Mike would. In the end (sit down before reading this) Mike agreed that there were tradeoffs to both approaches, and that there are longterm advantages that the Mac platform will have over Windows 95.
On the other hand, it will soon be easier to switch a Macintosh network over to Intel hardware, assuming some group of hardware vendors play by the rules of plug and play. Eight-character file names were probably a real deal-killer for wholesale conversion of Mac networks. When Windows 95 ships, it'll no longer be a barrier. Mac files will convert much more easily, in fact it's conceivable that ex-Mac users will be the most powerful Win95 users.
Anyway, next time a Windows versus Macintosh debate comes up, try pointing out the advantages of long file names, and hardware-based plug and play. The argument seems to work. If they say "but Apple is a total mess" (a common argument, used by Maples, of course -- I am paraphrasing tho), respond by saying it doesn't matter any more because if Apple doesn't want to play, you can always go down the street to Radius, or PowerComputing, or...
Speaking of warriors, Metrowerks, the original red-meat-eaters from Canada, have announced a deal with General Magic, that finally gives GM a non-empty developer story. It's a good move for Magic, and raises Metrowerks' visibilty substantially. I wrote about both General Magic and Metrowerks in the same piece, The DaveNet Letter, 2/13/95. Maybe both companies are reading DaveNet, and do quick deals, or maybe I just got lucky!
In any case, Metrowerks promises that soon we'll be able to develop Magic Cap apps using their Macintosh-based development system.
This is good news for Magic, because Metrowerks has a no-bullshit can-do reputation among Mac developers, and as Michael Miller, firstname.lastname@example.org, editor in chief of PC Magazine, points out, they saved Apple's ass in PowerPC Land. Without Metrowerks, he observed, there would have been no way to develop PowerPC native apps for the Mac. He's totally correct. How different Apple's world would be if Metrowerks didn't exist.
Maybe they can make the same difference for General Magic.
In the world of Mac developers, getting an endorsement from Metrowerks is like having Goldman Sachs take you public. You're definitely going to get attention. But if there isn't any substance to the story, if the development hurdles to Magic Cap are still insurmountable, it will turn into a negative for the company.
PS: What about Telescript? There's no mention of script development tools in their press release. Despite my complaints about OSA, my recommendation to GM is that they make Telescript an OSA-compatible interpreter, so Frontier and other Mac-based development environments can be used to develop and present Telescript scripts.
PPS: Re my description of a new kind of PDA in my last piece, Doc Searls, email@example.com said: "Find a precursor to this in Zenith Data Systems' new CruisePad. It's profiled in HotWired, under Fetishes, which I surfed to from the hot list at your Website. The guy who fathered it, Alan Soucy at ZDS, is a good guy; but the product is a bit of a stepchild for the company, and I don't think the product's getting much notice. FWIW, I helped name the product; but my first choice was Z-Cruise."
PPPS: As always, if you aren't interested in this kind of stuff, send me email and I'll happily delete your name from the list. And it's OK to forward it or repost it anywhere you like.