| d2r diego's weblog |
the web is not the browserYesterday Robert Scoble, reacting to comments by Dave, Jon, and others, was saying something really, really weird which involved comparing Microsoft with Dave. Robert's comment was centered on the idea that "Dave Winer has done more to get me to move away from the Web than a huge international corporation that's supposedly focused on killing the Web." Robert's thesis is summarized in the sentence that follows: "[...] what has gotten me to use the Web less and less lately [is] RSS 2.0." He goes on to describe the wonderful advantages of Longhorn's components, for example: "And wait until Mozilla's and other developers start exploiting things like WinFS to give you new features that display Internet-based information in whole new ways." Ok. On to the debunking. :) There are two points here. Number one, let's take Robert's thesis at face value. It is a standard Microsoft tactic to say "see? The little guy is doing it, so why can't we?". How MS can't see the difference between "the little guys" and themselves is beyond me. But that aside, there's a bigger (much bigger) problem. Robert: the web is not the browser. Robert says that he's "using the web less and less" because of RSS. He's completely, 100% wrong. RSS is not anti-web, RSS is the web at its best. The web is a complex system, an interconnection of open protocols that run on any operating system. Robert reads RSS on a Windows client (I assume), through a protocol originally developed in Geneva and now maintained in Cambridge, MA. He reads from a variety of servers, Linux, Windows, Apache, IIS, what have you. The RSS files he reads are generated by a multitude of software systems, all of them connected through the simplicity of a few lines of XML code. Now, if that is not "the web", then I don't know what "the web" is. Consider the alternative: what if Robert is right? What if in 2007 or whenever Longhorn leaves the realm of promiseware developers start switching in hordes to it? Take Robert's example: of "Mozilla's and other developers start exploiting things like WinFS to give you new features" and consider: what server will that run on? Linux? Not a chance. What if developers start using XAML? What client will that run on? Macintosh? A Symbian mobile phone? Of course not. You'll need a Windows device to see it. Longhorn is anti-web because it locks down everything back into Microsoft's control. It has nothing to do with HTML. It pushes a system where you'd be forced to using Microsoft servers and Microsoft clients. Let me say it again. The web is not the browser. The web is protocols and formats. Presentation is almost a side-effect. And that's what people like Dave and Jon are talking about. Categories: technologyPosted by diego on November 14 2003 at 12:18 PM Comments (please see the comments & trackback policy).
Robert Scoble uses Radio UserLand, so it's likely he reads RSS files presented as Web pages in Internet Explorer. Posted by: Rogers Cadenhead at November 14, 2003 3:50 PMI think I read at some point on his weblog that Scoble uses Newsgator as his aggregator. Otherwise it's hard to explain why he mentioned explicitly that he uses Radio for posting only. :) Posted by: Diego at November 14, 2003 5:01 PM
Remember WAP a few years ago was being advertised as the "Wireless Web"? Well, they've stopped doing that now because it wasn't true then and still isn't now. Now people say "Mobile Internet" which is much more accurate. If you swap out HTML and web browsers and the interconnecting links, you no longer are using The Web (sorry!). However, if you are use XML instead of HTML, then the name for http + xml has already been coined and that's "Web Services". It's an important caveat to append the word "services" because it's meant to imply that these things are not supposed to be used directly by humans. RSS, XML-RPC, SOAP, etc. all use http but *have no standards* when it comes to the UI presented to the user nor the links that may be contained within. They are not The Web. Okay, that said, Microsoft has been pushing the idea of Web Services for many years and will continue to do so. They want to kill The Web as we know it and replace the open-speced display and user interface technology called HTML with custom and proprietary XML Web Services which render to "rich clients" instead. Whereas HTML can be rendered on a variety of platforms and browsers and is the real threat to Microsoft's dominance, WAML and all these other specs come right out of Microsoft's proprietary playbook. Define a new "open" spec, get people to write to it, make it impossible for other platforms to duplicate the experience, lock in the users. -Russ
Actually, Scoble uses NewsGator, which runs in Outlook. Also, I think Diego and Mark Baker are missing the point when they say "the web is not the browser". Scoble is reacting to the past couple of weeks of XAML-bashing where conspiracy theorists are claiming that XAML is an attempt to kill HTML, and by extension "the web". Scoble is just pointing out that "rich client" is a trend that is moving forward with or without MSFT. Posted by: Joshua Allen at November 14, 2003 6:11 PMYes, this MS policy is everwhere. The company I was working for a while back was looking at using MS project to do our project maintenance. It's a slick program, you can schedule resources, have access to modify and update through email and web, all in a slick and user friendly interface (note: I've never used it myself beyond seeing it, so I don't know how useful it actually is :). We were a mostly linux shop, developing a linux firewall product. All our servers were linux machines, as were our firewalls. Developers used linux to do the C/perl/java that was needed, sales and execs used windows for the word/ppt/etc they needed. The problem was that to have the cool features available for project you needed to have an IIS webserver. And an exchange server. I understand that MS isn't going to develop for apache (their "competition"), but this sort of made the party line quite clear for me. They were saying that to get full use of the product you had to invest in another $N of their software and support. Of course, if you're going to run exchange for just this, why not replace that silly unix mailserver, and that apache web server thingy as well. We did not. I can see other companies and PHBs doing this though. Posted by: Arcterex at November 14, 2003 6:59 PM1) "Wireless Web" is a registered service mark of Sprint PCS for their WAP-over-IS95 service, which since mid-2002 is being phased out in favor of HTTP-over-Mobile-IP-over-1xRTT, which they call "Vision". The browsers on current Sprint (and Verizon, AFAIK) phones grok real, honest-to-God HTML (although, admittedly, only in bite-size chunks). The usability problems of wireless handsets have a lot more to do with the total lack of a practical means of text entry than they do with the protocols. It's certainly valid to say that WAP is not Web, in the same way that Gopher was not Web, but that doesn't make it evil or bad. 2) Practically every large corporation I have ever worked for (several at this point) has used Exchange for internal email. Exchange is almost as ubiquitous as Office or Project (to which, as far as I can see after twenty-odd years of GANTT charting, there is no practical alternative). So it's not unreasonable to assume that Project customers will be using Exchange, and will be at least open to the possibility of IIS. Protocols are not ends in themselves; protocols are means of delivering features to the user. If you don't like being held prisoner by vendor X because you're addicted to features, the thing to do is not to complain about vendor X, who brought you the features in the first place. The thing to do is to clone the features, or convince vendor Y to do so. Russ, RSS does use HTML, this is the center of much bickering on the mail lists, but it's common practice to encode the HTML the user enters in the description sub-element of item. Purists say this is evil, but it's a fact of life. In many if not most cases, RSS is a structure wrapped around bits of HTML. It's interesting to note that Atom has failed to resolve this issue. The faceoff continues. To me it was never a question, it's a user-driven thing. If you strip the markup as the text is flowing out to RSS the users call it a bug. Are they right? Hey remember what they say about customers. To Diego, it's Microsoft's dream to turn the clock back to 1993, the end of their brief period of total and utter domination of the computing world. That kind of hegemony is unnatural, and quickly gets defeated. Something new and hot will pop up just as Longhorn is about to wrap it all up, and make all their innovating (ie megalomania) totally irrelevant. Posted by: Dave Winer at November 16, 2003 2:05 AMOkay, where do I start? Russ: you say the web is HTTP + HTML. Okay. However, what you type in your browser is a URL. The URL is post-web (RFC 1738 is dated Dec. 1994), and as it says in the RFC: "The specification is derived from concepts introduced by the World-Wide Web global information initiative". Yes, originally "The Web" was the world wide web, HTML+HTTP. But very quickly things became intermixed. Today, you'll click on a link that is "ftp://" to download a file. FTP dates back to the pre-WWW days. Is clicking on ftp:// *not* the web too? If people click on a text file obtained through FTP and read it on their browser, is that not the web? Or on a quicktime video? Today we often use HTTP to download files, not to see HTML, or to stream video, or audio, or Flash animations, or, yes, RSS, and it's an integral part of the "web experience". Some clients take HTML and present it in different ways "syndicating" it. So what I'm saying is that, even though at the beginning the web was indeed just HTTP+HTML, it went beyond that very quickly. Even if you only consider what you see within the browser the "web experience" includes a multitude of formats and protocols. Arcterex: that sounds like a relatively typical situation from what I've read. Microsoft has been integrating "vertically" more and more, particularly after Windows 2000. Jeff: I don't see what a trademark of AT&T has to do with anything, but ok. :) However, on to point 2), you say: "Protocols are not ends in themselves; protocols are means of delivering features to the user. If you don't like being held prisoner by vendor X because you're addicted to features, the thing to do is not to complain about vendor X, who brought you the features in the first place. The thing to do is to clone the features, or convince vendor Y to do so." Well, the small problem is that if the protocol is not open for others to implement (or, if it is the case with an API like Win32, it keeps changing constantly without everything documented) no one will be able to follow it. However, assume that they could be followed, with a lot of hard work and a considerable "time lag", this assumes that the protocols have a chance to fight it out in an open market, and "the best wins". But it's difficult to see how that can be the case when a company like Microsoft has so much market power that it can, for example, deploy tens of millions of clients within a couple of years. It's not as easy as "ask vendor X to implement it". Who would be "vendor X" in this case? Dave: I agree with you that they are trying to get back to a 1993-like situation ("-like" because in 1993 there were lots of credible alternatives to the whole stack, OSes, office apps, servers, etc, and today there are only very few). I do think however that they are a bit more open now and that they are trying to engage in a dialogue. I have no idea if that matters one iota for the intended final result :), I guess we'll have to see. Diego: I think the history of the industry bears me out when I say that software is perishable: you can only sell a given application for money for so long before someone clones it and gives it away for free, or at least undercuts your price so far that you can no longer turn the profit to which you have become accustomed. There are a few exceptions to this rule (e.g., Mathematica, and to some extent AutoCAD), but they're growing fewer all the time. Samba, WinE, Gaim, and OpenOffice prove that Microsoft is just as vulnerable on this score as anyone else. Last month Microsoft moved to reassert access control to their instant message service (the server side of which, btw, is part of Exchange 2000) by forcing all users to upgrade clients and attempting to license the new protocol to independent developers; within a week of the cutover, third-party clients had hacked the protocol changes and were back online. It's worth noting that SMB, Office, and MSN Messenger have already been cloned, and Exchange and Project haven't. By and large, the little guy - the guy who can't afford to hire a Brooks Brothers-clad, Wharton-MBA CIO and give him an eight-figure budget - doesn't need things like Exchange and Project. Should I need to log in to somebody's VPN and check email on an Exchange server, I can do that with my Mac :^). Posted by: Jeff Carroll at November 17, 2003 1:19 AMCopyright © Diego Doval 2002-2007.
|
