Welcome here
What led to the creation of SGML?
Before SGML there was IBM's GML—not standardized. It was created in 1969 to solve what today we would call an application integration problem. I was trying to get three separate programs to work together to form a law office information system—an editor, a formatter, and an information retrieval program. I hit on generalized markup as the means to let them share data. With Ed Mosher and Ray Lorie, I developed that idea into GML.
I continued research on document structures and created additional concepts, notably the validating parser in 1974. At that point, SGML was born, although it took another decade before it was fully developed and standardized.
What would you have done differently?
I'd have documented it better, especially the use of notations for strong datatyping and the relationship between documents and data.
How will we reach the goal of truly separating presentation from data?
We've seen it happen in most applications of XML because XML is used chiefly for data interchange and messaging. Those areas emphasize the processing of the abstract data, so rendition is normally handled with a separate style sheet. But browsers and traditional rendered Web pages are, of course, another story. I think we'll start to see an improvement as we ship more XML directly to the desktop, rather than converting it to HTML on the server. Now that the major browsers can support XML, that trend will increase.
Content management systems are also in their early phases and known for their cantankerousness. What will the future hold for such systems?
As enterprises increasingly recognize the importance of content management, content management systems will be layered on DBMSs and application servers. They're likely to spur the growth of native XML DBMSs and XML layers on relational DBMSs.
Robert Hopkins
Founder, Weblations
What's wrong with automatic translation tools?
Machine translation (MT) tools simply don't work very well. To sum up recent progress somewhat cynically, I would say that today you can get a bad machine translation in only a fraction of the time it took 15 years ago.
MT will someday be improved through new algorithms based on the latest research into how the brain processes language. Meanwhile, computing power is increasing. Researchers are building huge databases of translated material to feed and test the new algorithms.
How is Unicode affecting the World Wide Web? Is it making multilingual Web sites easier to build and maintain?
My company specializes in Web site localization. The Unicode revolution has brought huge benefits to our tools and platforms, but strangely, it's non-existent on the Web itself.
When you browse a page on the Web in English, Chinese, or whatever, the chance that it's delivered in Unicode or UTF-8 (the Internet's compact version of Unicode) is less than one in a million, literally.
What's the biggest obstacle to Internationalization, and how will it be overcome?
The biggest obstacle to internationalization is the sheer distance, in every sense of the word, between the content owner and the content user. The people who benefit from I18N mostly are not Americans, while those who pay for it mostly are.
Most of the biggest knowledge bases are available in English only. Why? I guess it's because the Americans with the big budgets don't understand in their bones how absolutely imperative it is to translate that material to get a return on it outside of the U.S.A.
Richard Luna
CTO, PasswordHeaven.com
What's the number one security problem with the Internet, and how will we fix it?
A lot of areas are susceptible to attack, but one area touches most of us: email. For all of its benefits, email is clear text. The vulnerability is compounded by TCP/IP.
The answer is encryption. Some email is encrypted today, but unfortunately most devices can't access the encrypted email, and therefore, the standard is a non-standard.
Microsoft has said it is rededicating itself to security. Will it succeed?
No. Microsoft's primary focus is increasing its dominance in the market. Strengthening security doesn't help in that area.
One example is what it did with the XP TCP/IP stack. In previous versions of Windows, the core IP address couldn't be spoofed. That meant that if an attacker was using a Windows-based PC, the attack could be traced back. With XP, Microsoft changed the stack so that the originating IP address can be spoofed, making XP a good hacker PC. If Microsoft were serious about security, it would have left the TCP/IP stack alone.
Is biometrics something to watch?
For biometrics to be beneficial, it has to work across all operating systems, not just one or two. Market forces will prevent that from occurring.
Would Microsoft ever let a mainframe company or Sun dictate its authentication? All of these vendors know that owning the authentication has the effect of owning the enterprise. All of them, from Novell to Microsoft, Sun, and IBM, are trying to own the enterprise. Because there are so many vendors competing for authentication ownership, no one group will win, and therefore biometrics will be relegated to small or platform-restricted applications.
Also, the authenticating system has to have a pattern to match with. Without the original pattern, the authentication mechanism fails. With biometrics, the authenticating pattern is your fingerprint, palm print, retina, etc. I don't believe end users will give their consent to store their authentication information.
What do you see on the horizon for encryption?
Encryption is the future. Whether we're talking about VPN communications or email, encryption is the only mechanism that makes Web use truly secure.
As more systems are interconnected, encryption will be required to keep the communication lines from being disrupted. That said, before encryption gets fully deployed, I think—unfortunately—we will see several major disasters. Just as the September 11 attack on the World Trade Center reminded the IT world not to concentrate technology in one location, managers will need to experience an attack [before they will] provide financial resources for encryption.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment