Cringely adds to the XML antiFUD

From: Karl Anderson (kra@monkey.org)
Date: Fri Apr 13 2001 - 11:34:32 PDT


I found the latest Pulpit to be pretty annoying.

http://www.pbs.org/cringely/pulpit/pulpit20010412.html

  The power of XML, then, is that it makes applications aware of what
  they are about. An XML search engine, for example, wouldn't have to
  drag back all the text and analyze it for content. It would just
  send out a message saying "All pages that are about fly fishing,
  please identify yourselves!" And they would.

eh? He makes a little more sense later when he says that XML's value
comes when it's controlled and driven by a taskmaster like .NET.

  Data, Know Thyself The Power of XML is Going to Change Everything
  About Computing. Now If Only I Could Describe It

  By Robert X. Cringely

  I've been waiting to comment on Microsoft's .NET business because it
  seems to me that .NET is far less important than the XML technology
  that lies beneath it. Same for Sun ONE, the most obvious competitor
  for .NET. XML is more interesting. But there is a risk in tackling
  such a technical subject. The risk is that my effort will be
  unsatisfying to nearly everyone. To the experts who breathe XML,
  what I write will inevitably look naive and maybe even incorrect. To
  the lay readers of this column, it might read like gibberish. So
  I've lain this week in a fetal position and tried to connect
  emotionally with both groups. What follows is my take on the real
  meaning of XML. If you think I'm stupid, well maybe so. Just don't
  tell my Mom.

  XML is the new religion at Microsoft. CEO Steve Ballmer calls what
  is happening "the XML revolution" -- a revolution so important that
  Ballmer is working hard pivoting all of Microsoft to take advantage
  of it. In three years, he says, XML is going to change completely
  the way we use computers, and most of Microsoft's competitors see it
  that way, too. But what the heck is XML, anyway, and why should we
  care?

  XML stands for eXtensible Markup Language, and is just the latest
  descendant of the General Markup language invented years ago at IBM
  as a kind of common style sheet for technical reports. The General
  Markup Language begat the Standard General Markup Language that
  begat the HyperText Markup Language (HTML) that made possible the
  World Wide Web. All these predecessors to XML were about describing
  a page, whether on paper or on a computer screen. Where is the text?
  Where are the pictures? What fonts are used on what color
  background? When a web page is written in HTML, it can be rendered
  (drawn on the computer screen) by any browser application like
  Microsoft Internet Explorer or Netscape Navigator running on almost
  any kind of computer, and should look pretty much the same on each,
  right down to those clever push buttons, check boxes, and radio
  dials.

  What's odd about HTML is that while it does a perfectly good job of
  describing how a web page will look and function and what words it
  will contain, HTML has no idea at all what any of those words
  actually mean. The page could be a baby food ad or plans to build an
  atomic bomb. While HTML knows a great deal about words, it knows
  nothing at all about information.

  Say we want to find specific information somewhere on the
  Internet. Most of us would use a search engine. Like every other web
  page, a search engine like Excite or Google is presented using HTML,
  but it isn't the HTML that finds us the stuff for which we are
  looking. HTML is not a programming language. People don't write
  programs in HTML. Instead, HTML describes how a web page looks and
  in that capacity it contains the text of that page, but to HTML,
  words are just words, not information. It is up to the search engine
  to invoke another program to first parse (read) the HTML, stripping
  away all the parts having to do with page description and leaving
  just the text. Then the text has to be identified and read by the
  search engine program, which is generally looking for key words
  specified by the user. It's a dumb system that usually only looks
  for word patterns in the text with little regard for what the text
  is actually about.

  HTML describes web pages but has nothing at all to say about what
  the page means. It describes fonts and positioning and which parts
  of the page are hyperlinks and which are text, but the HTML doesn't
  know whether the page is about fly fishing or pornography. XML, on
  the other hand, describes data, not pages. It is all about fly
  fishing or pornography -- about the actual information content --
  but says nothing about layout. You still need HTML for the layout
  part. So XML is not in any way a replacement for HTML.

  The power of XML, then, is that it makes applications aware of what
  they are about. An XML search engine, for example, wouldn't have to
  drag back all the text and analyze it for content. It would just
  send out a message saying "All pages that are about fly fishing,
  please identify yourselves!" And they would.

  XML makes web content intelligent. And by doing so, it enables us to
  move beyond the current world where we look at the Internet through
  browsers to a more advanced world where every application is
  Internet-aware and maybe the browser disappears as a popular
  application. Once your spreadsheet talks XML, it can link across the
  Net into other spreadsheets and into server-based applications that
  offer even greater power.

  That's at the heart of Microsoft's .NET (dot-NET) initiative, which
  puts little XML stub applications on your PC that don't actually do
  much until they are linked to the big XML servers Microsoft will be
  running over the Internet. All your office applications become
  XML-aware, which means you can do powerful things on tiny computers
  as long as you continue to pay rent to Microsoft. The effect of
  dot-NET is cooperative computing, but the real intent is to smooth
  Microsoft's cash flow and make it more deterministic. .NET will move
  us from being owners to renters of software and will end, Microsoft
  hopes forever, the tyranny of having to introduce new versions of
  products just to get more revenue from users. Under .NET, we'll pay
  over and over not just for the application parts that run on
  Microsoft computers, not ours, but we'll also pay for data, itself,
  with Microsoft taking a cut, of course.

  Probably the biggest reason why we even care about this stuff is
  because of the Y2K computer crisis of the late 1990s. Back then, we
  had hundreds of thousands of mainframe computer programs that didn't
  make a lick of sense to even the smartest programmer trying to read
  their code. Their original programmers had retired or died, taking
  with them any notion of what most of the computer code actually
  meant. So a whole new batch of programmers had to find ways to
  assign meaning to what appeared to be gibberish. What made the job
  so difficult was that the code wasn't readable by humans. Huge
  databases contained hundreds of millions of entries, but which words
  were the customer names and which where their account numbers? It
  wasn't at all clear, and every program was different from every
  other program.

  XML changes all that by introducing the concept of metadata -- data
  about data. In XML, each piece of data not only includes the data
  itself, but also a description of the data, what it means. Now your
  XML database can have a list of names (that's the data) and a tag on
  the data saying that these are customer names (that's the
  metadata). Should some Y2K-like catastrophe afflict our XML
  database, it would be easy for any programmer to look at the
  metadata to reconstruct the database program. In fact the metadata
  is the program, which is how those fly fishing pages were able to
  announce themselves in an earlier example.

  Once we embrace XML, and nearly the entire computer industry already
  has, then wondrous things begin to happen. Airline ticket databases
  suddenly are aware that's what they are. So within the constraints
  of a vocabulary limited to words like "passenger" and "seat number,"
  finding the cheapest way from here to there becomes a matter of just
  asking. The query -- the question you are trying to answer by
  analyzing data -- becomes the database, itself.

  XML is leading to a fundamental shift in the way that people store
  information -- away from the traditional use of documents and files
  to thinking about a document as a container into which various
  pieces of data are combined or "poured" as required. This kind of
  document, which is the typical product of an XML application, may be
  stored for an extended period of time or it may exist only as long
  as a person interacts with it.

  With XML, where the information resides becomes less important than
  being able to access it. This leads to cooperative applications
  where modest client machines like handheld computers or mobile
  telephones can use powerful servers to accomplish major computing
  tasks. Imagine a television reporter who shoots video of an event,
  uploads the video to a server, then edits the video using only a
  mobile telephone. XML can make that possible.

  But wait, there's more. That "X" in XML stands for "extensible,"
  which means this is a language to which you are allowed to add new
  words. And what those new words can signify is almost unlimited. Now
  that XML makes data more or less self-aware, it is possible to have
  that data announce itself as it changes. When a favorite stock
  price jumps or falls, you'll know, because an XML application tracks
  that stock which now squeals out its state of being trade-by-trade.

  The end result of adding XML throughout the Internet will be a
  change in web infrastructure. We'll put much more effort into
  maintaining and updating data and much less effort into presenting
  it. Today, huge server farms are used to present to users over and
  over again what can be beautiful but essentially stagnant pages of
  stale information. Tomorrow your computer, whether it is on your
  desk of your wrist, will directly query XML data sources to generate
  dynamically not the web page as its authors want you to see it, but
  exactly the web page you want to see. And this page will use only
  up-to-the-moment data -- data that you will probably pay for. And
  there lies the business opportunity, providing these new data
  services.

  Who will lead this new business? Though XML is an open standard,
  usable by anyone for free, Microsoft is the early favorite for
  commercialization with its .NET platform. One aspect of Microsoft's
  XML push that has gone generally unnoticed is the role of Great
  Plains Software, Microsoft's most recent and second-largest ever
  acquisition. Buying Great Plains, a maker of financial software for
  small and medium-sized businesses, has everything to do with
  .NET. The Great Plains customer base is ideal for Microsoft's new
  web services.

  Going head-to-head with Microsoft, as always, is Sun Microsystems,
  whose XML platform is called Sun ONE. Like Microsoft, Sun's offering
  is intended to attract companies that want to provide XML-based data
  services. But Microsoft and Sun have plenty of competition in this
  emerging market. All of the leading tech vendors, including IBM,
  Oracle, Sun and BEA Systems, are pushing their own technologies for
  individuals and companies interested in writing and running web
  services.

  If there is one company that will benefit no matter what, it is
  Cisco Systems, because one characteristic of XML is data
  expansion. Putting data in XML format and including the metadata
  description inevitably makes data bigger because we are shipping
  over the line not just the raw data, but also a description of the
  data. Bigger data slows down the Internet. A slower Internet makes
  network service providers build bigger network connections.

  And newer, bigger network connections always means more work for
  Cisco.

-- 
Karl Anderson      kra@monkey.org           http://www.monkey.org/~kra/



This archive was generated by hypermail 2b29 : Sun Apr 29 2001 - 20:25:42 PDT