Copyright © 1995-2004 UKUUG Ltd

UKUUG

Newsletter Section 5

Across the Pond

Bob Denny's 1997 Predictions for the Internet

WebSite developer Robert Denny, whose software products have been used by more than 100,000 people worldwide, has made 11 predictions for 1997. Denny is well-known as the developer of the award-winning Web servers WebSite and WebSite Professional. The predictions are:-

1. The intranet will not “take off” in 1997. Its presence in corporate America will continue to increase at a gradual rate over the next 3-5 years.

2. There will be a big shakeout among Internet Service Providers (ISPs). Those that will survive will fall into two categories: either they will be among the largest, or they will have value-added services that are reliable and cheap.

3. Microsoft will make inroads into the browser market, but won't knock Netscape out of the picture.

4. There will not be any serious collapse of the Internet, but it may get more sluggish, more often.

5. Getting on the Internet will remain cheap, getting down to the bare-metal pricing – and it will be flat-rate. But...

6. Microsoft will reveal plans to buck the Internet flat-rate trend with some form of usage-based pricing for their products. There will be lots of controversy surrounding this.

7. Java will fail to grow into a broadly accepted technology.

8. Cable modems will not become popular, despite last year's hype. Their numbers will remain small as a percentage of total browsers.

9. ActiveX in Web pages will be a lot less important than Microsoft would like.

10. For newcomers to the Net, live chat will become much more popular.

11. The sleeper technology: in 1997, someone will really figure out MBONE (multicast backbone) and this terrific technology will take off.

Denny noted, “The future looks less clear now than it has looked in the past 10 years. Many technologies have been floated out there – a lot of them are half-finished but the technoids are jumping on them. The big question is: will the de facto standards that have made the Internet what it is today, that have given it strength, endurance, refinement and stability – will they be replaced and to what degree by 'standards du jour' (when a big company says there is a standard but it is not in general use)? If that happens, it's possible that those qualities will be destroyed, and replaced by a techno-political structure that inhibits creativity at the infrastructure level.”

Denny elaborated on some of his predictions as follows.

1. Most companies that are big enough to have real intranets amortize computers

over 5 years; the average age of their computers is 2-3 years. So companies are not making large investments in the intranet now. In the end they will, but it will be over a multi-year time frame, 3-5 years.

2. The ISP shakeout will not be one of consolidation, as many will go out of business. The key to capturing the business is flat rate. Even America Online has adopted flat rate pricing, which will hurt the business of ISPs.

4. There won't be a dramatic collapse of the Internet. It's too vibrant, too alive, and there are too many people working on the problem. So for the short-term, the Internet's capabilities look strong. Within the next few years, it may become more difficult to add publicly visible nodes.

5. Pricing is low now, and there is every reason to assume it will stay there (see #2 above).

6. Microsoft has already started down the road of user-based pricing in several ways. For example, they are promoting pricing for servers by saying, in essence, “if you want a Microsoft server that supports X number of users, you have to buy the higher priced platform that supports a server for that number of users.” And their new Denali active server creates the notion of Web 'sessions', which could result in charges for simultaneous 'sessions', 'sessions' per unit time, or total number of 'sessions': a postage meter concept. But the Web is not user-based, it's transaction-based, so it doesn't make sense to try to charge for services per user. If Microsoft actually does succeed in getting people to accept user-based pricing for Internet services, that could reduce innovation by a huge amount.

7. Java's capabilities in the browser area are stunted. This happened because it was developed at Sun, and then Sun licensed Microsoft to do the core reference implementation for Win32. But Microsoft wants people to use Visual Basic and Component Object Model (COM) implementation language, which is proprietary. This will limit Java's growth. One interesting question here is: where will Microsoft make its money, following their extensive investments in technologies such as Java?

9. When someone really sits down to build a Web page, they have to ask: will I restrict this to Microsoft's proprietary system? What about people with UNIX machines, Macs, etc.? I don't think people will be willing to risk having a site that only Microsoft Internet Explorer users can view.

Denny sums up his thoughts on what he considers to be the most pressing issue facing the Internet today: “It ain't a done deal that Microsoft has cornered the Internet, and there's a strong chance that they won't. This means that proprietary Microsoft-only technology won't get nearly the deployment some people might expect, with the exception of closed, intranet situations where they're willing to go 'all Microsoft'. Those who choose this route may automatically cut themselves off from the outside world, should they ever decide to use their internal information structure to communicate with users, vendors, etc. That would be a reprise of what happened with companies that chose all-IBM, and I hope we've learned from that.”

USENIX Conference on Domain-Specific Languages (DSL)

Important Dates for Refereed Paper Submissions
Papers due : 13 June 1997
Author notification: 10 July 1997
Camera-ready final papers due: 2 September 1997

Language is central to the discipline of software engineering. Programmers use a variety of languages in their daily work, and new languages appear frequently. This proliferation is not gratuitous: each new language offers specific solutions to genuine software problems. Domain-specific languages (DSLs) are explicitly designed to cover only a narrow class of problems, while offering compelling advantages within that class.

This conference is dedicated to the discussion of the unique aspects of DSL design, DSL implementation, and the use of DSLs in software engineering. The conference seeks to advance the practice of DSL design, DSL implementation, and software engineering generally by:

.     eliciting examples of successful domain-specific languages

.     highlighting the spectrum of benefits which domain-specific languages can provide

.     discovering design principles and methodologies for creating DSLs

.     eliciting design techniques and tools for working with domain-specific languages throughout the software engineering lifecycle

.     providing a framework within which language designers from different domains can easily communicate

.     establishing the practical value of domain-specific languages through the publication of empirical data concerning productivity, quality, and maintainability

.     creating a community that will continue to study and refine the practice of software engineering through domain-specific languages

There will a one-day tutorial program, followed by three days of technical sessions. The technical sessions will offer refereed papers, invited talks, and Birds-of-a-Feather (BoF) sessions.

Submission Instructions

See the conference web page URL: http://www.usenix.org/wits97/ . Or email to info@usenix.org ; in the body of your message, state “send wits97 conference.”

Registration materials will be available in August 1997.