by David Strom

(for Connextions newsletter, 3/96)

Writing about the Web for a print publication is quite a challenge: things change so fast that sometimes one's research is out of date not only before publication but before one's article is even saved on the word processor.

Our esteemed editor asked me to do an article looking at home pages in the style of Siskel and Ebert (I paraphrase his request). I'm not so sure what this means, but I will attempt to give you a brief look at where things are going in terms of web developments from my perspective as well as some of the things I'd like to see happen with the web over the next few years.

First, my top ten list of web trends.

1. Caffeine and anti-caffeine. For purposes of discussion, let's call all efforts to combine the web with multimedia, animation, agents and complex images under the heading of caffeine, and other efforts to combine the web with proper search engines, directories of links, and mostly text-based methods under the heading of anti-caffeine. Both "sides" if you will are important for the web to grow and thrive. But developers for the most part have chosen one direction or another for web-based products. Most of the trade press has focused their attention on the caffeinated side of things, which is a Bad Thing: I believe the more interesting developments will be on the anti-caffeine side. But perhaps I am biased: I also prefer decaffeinated beverages too.

   Nevertheless, these caffeinated applications are important, if nothing because they free us from depending on a particular computing platform and also allow for more network-centric computing and program development. Now, do I believe this will happen anytime soon? Nope.

2. Netscape and anti-Netscape. The general public loves a two-sided battle, and one of the problems was that until recently the web had so many sides that it was difficult to keep track. Now, thanks to Netscape, it is much easier: we have Netscape vs. Everyone Else, or Netscape vs. Microsoft. (You can fill in your favorite vendor if you'd rather.)

   Netscape has had some positive effects on the web: it has brought about a rapid deployment of graphical (versus text-based) browsers, innovation with respect to HTML tags and servers, and improved graphical look and feel of web pages themselves ("this page has been optimized for ..."). It has also brought about some negatives as well: a rapid deployment of overwrought graphical content, too frequent updates of browser software, and too many new tags and server extensions. All of this contributes to what I feel is the end of openness and a standards-based web[1], which is a Bad Thing.

3. Five million channels and nothing on. First everyone on the Internet had their own (potential) home page. Then came along Compuserve, AOL and Prodigy and now those several million customers now have their home pages too. Soon everyone will have their own web server. Then what? Actually, this is a Good Thing: I believe the best role for the web is to have a web server into each desktop operating system, and by proliferating home pages (whatever they really are these days is hard to tell, but that's for another article) only helps to get this notion going. Why is this desirable? The kind of things that a web server does (linking and organizing documents) is exactly the kind of things that a modern operating system should do, but doesn't -- yet.

   An interesting corollary to this trend is that everyone becomes their own publisher. This is not necessarily a Bad Thing. Take my own case as a shameless example. I publish my own email newsletter now to a small audience of computer vendor marketing and engineering types. While I could do this before the rise of the web, having a web site to archive my back issues and provide links to other interesting places complements the overall publishing scheme nicely. The trick to being your own quality publisher is also finding the right mix of skills (editorial, production, sales and circulation) that are needed for the world of paper publishing as well.

4. NT vs. Unix. The web has been a great example demonstrating the best and worst of Unix: if you know Unix, setting up a web server is no great stretch. The tools to administer a web server have great power but obscure commands, and to the Unix-ignorant are even more intimidating than the basic operating system. All fodder for the Windows NT camp. The web and NT make beautiful music together: you can have your graphical user interface cake (something that all Windows users have taken for granted but that Unix users only recently discovered) and eat reliable, rock-solid operations too (something that all Unix users have taken for granted but all Windows users have only recently discovered).

   NT web servers are also for those corporations that have relative control over their desktops and are comfortable with (pre-NT versions of) Windows. There are now over a dozen NT web servers[2], and I'm sure more are on their way. But more importantly than sheer numbers is the perception that NT makes a great web server by the corporate buying public. My recommendation: if you don't know Unix, don't start now for all things web. Try one of the more popular NT web servers.

   But don't think that an NT box can completely replace all the functionality of a good Unix server for all things non-web (mail, network management, and net news come to mind as three services that Unix does particular well and NT is still far short on). In addition, some signs of immaturity still remain: NT web server remote administration tools are bare-bones at best. Too few Internet Service Providers offer NT web hosting services, although that will change quickly over time. And sometimes it can be confounding to have too many choices for software.

5. New class of useless software: HTML editors and HTML add-ons to existing word processors. The web has brought about the fast demise of proprietary desktop word processing formats in favor of tagged text. Is this progress? A brief history of word processing, for those that might have forgotten: Back in the old (pre-PC) days, we had VT-100 terminals and control characters. TeX [3] if not king, was certainly a notable achievement in this venue. Then came Wang, NBI, Vydec, Xerox and other dedicated machines that did the job: you all know what happened to them. Out of Wang's popularity was born Multimate, which ran on PC DOS. They got acquired by Ashton Tate which was acquired by Borland, and then disappeared. Now we have three major vendors of word processors on the desktop.

   Anyway, most of these HTML editors and HTML add-ons (Microsoft Word Internet Assistant, Lotus Word Pro, and WordPerfect's Internet Publisher) don't really add much on to the process of creating, checking, and publishing HTML documents. My recommendation: find yourself a good text editor and return to the glory days of the past.

6. Latest buzz word: the "Intranet." We now have Attachmate saying they are "the Intranet company," a trade magazine with an entire "Intranet" section [4], and I'm sure trade shows with Intranet in their titles aren't far behind. In one week last fall, Microsoft, Sun, Netscape, and others made product announcements with the word figuring prominently. I've even written a white paper on the subject [5]. So what is Intranet, anyway? My take is simple: take one part TCP/IP, one part Internet technologies such as mail and web, and one part corporate database, shake into a publishing metaphor and stir lightly into an HTML browser.

7. WEBng: database connectivity. Despite my attempts at humor, Intranets will become serious business as corporations figure out client/server applications really mean access to data. The Internet technologies were nicely designed for this purpose, and expect to see more corporations co-opting this mix as they get further along in publishing more than just employee handbooks and corporate policy manuals on their internal web servers. The ability to tie your mainframe and network databases and web servers together will make this a viable business in years to come. We are just beginning to see products that make this possible, and again this is another Good Thing.

8. HTML as the new application interface. We now have a variety of "faceless" applications -- software that has no interface of its own, but rather uses a web browser instead to do user input and display information. Want to manage your network laser printers? [6] Use a web browser. Want to check your calendar? [7] Use a web browser. Want to manage your router? [8] Use a browser. I believe this is a Good Thing, but not because I want to run my web browser for all my applications. Rather, this is desirable because it furthers the adoption of Internet technologies within the general-purpose office productivity market, and also makes it easier for me as a mobile office worker to move around the world and do my work over the Internet.

   What will really help this along is TCP/IP. TCP/IP is now available as part of every new desktop operating system sold today. That's a big change -- and a big improvement -- from several years ago, when the protocol was only the provence of the truly enlightened. But we still have several potholes along the IP highway to fill in: management tools, for example, that were designed for routers and not desktops; mixing earlier versions of Windows and IP is still somewhat of a challenge; and making Winsock and Open Transport work is full employment for an entire army of consultants.

9. Two words: Internet commerce. Depending on whom you talk to, either the Internet will never be safe for conducting commerce or is already is far and away safer than handing your happy waiter your credit card. Dick Shaffer of Technologic Partners sums it up best by saying "A standard payment mechanism for Net purchases will appear in 1996 (it will be called a 'credit card'). In the meantime, the various Internet *.cash systems will continue to duke it out.

   I got real insight into some of this process when I tried to buy some stuff via the Internet: shopping malls were confusing (not to mention relatively empty of patrons), incompatible forms with my browser, difficult to find stuff, overpriced shipping charges, etc. [9] All in all, a relatively unsatisfactory series of experiences. But this will get better, or else we will have to find some other use for the web fast.

10. Pipes vs. content: the latest tectonic shift for ISPs. Something I've predicted a long time is that the traditional on-line access vendors (Compuserve et al.) are moving to becoming providers of pipes rather than content. [10]. I don't mean to minimize the value of content on the on-line vendors at all, it is just that the Internet will receive more innovation and more interest. Already, AT&T and Apple have learned this lesson with Interchange and eWorld, respectively. This trend will continue, and we might even see phone companies figure out the former (but doubtful of the latter).

    So, what would I like to see happen on, and around the web?

    I'd like to see Microsoft stop talking about how good it is going to be and really try to integrate its web services into NT.

    I'd like to see true 32-bit applications appear on Windows 95 and NT so that I didn't have to continually be reminded of the 8.3 character file names inherited from a 20-year old operating system.

    I'd like to see authoring and publishing tools work well with a wide range of web servers so I didn't have to manually arrange my content and use an ftp client as my main organizational tool.

    I'd like to see a meaningful and stable HTML standard that was fully embraced by Netscape and Microsoft, so we can innovate in more useful ways than tag fights.

    I'd like to see the Macintosh reborn as a web authoring platform, with all the necessary tools and integration, and ease of use that it had as a desktop publishing platform.

    I'd like to see the trade press focus on something other than caffeinated applets and agents, and examine the lack of meaningful ways to transport existing content onto a web server.

    And to put this all in the proper perspective, I'd like to see World peace in my lifetime.

    BIO:

    David Strom is president of his own consulting firm in Port Washington, NY, where he works for leading-edge vendors of networking and communications products. In between consulting gigs, he is publisher, editor, and sole writer of Web Informant, a free-for-the-asking HTML-based newsletter distributed via email about web marketing for the computer industry. His web site, http://www.strom.com, contains back issues along with the usual links to interesting places on the web that he wish he thought of first. Strom's background is as an editor and writer in the computer trade press, where he wrote one of the worst articles on the InteropNet ever published. He was founding editor-in-chief of Network Computing (the US CMP version) and presently writes for Infoworld and Forbes ASAP as well as appears occasionally on CBS TV news talking about the Internet, but don't hold any of that against him.

    References:

    [1] A more complete essay by the author on the end of openness, "Anybody Remember Open Systems On The Net?" can be found at CMP Techweb

    [2] Check out our Webcompare list of web servers (along with a description of features).

    [3] TeX and METAFONT, Donald E. Knuth, Digital Press/American Mathematical Society, 1979. For places around the web that offer the TEX product.

    [4] Interactive Week magazine not only has an Intranet section, but an Intranet columnist.

    [5] Creating Private Intranets: Challenges and Prospects for IS by the author is available here.

    [6] Tektronix PhaserShare printer management software.

    [7] Campbell's OnTime calendaring software.

    [8] Tribe's WebManage routers.

    [9] An article by the author on the on-line shopping experience can be found in Web Review magazine.

    [10] See "In the On-Line Market, The Name of the Game is Internet," New York Times, 9/25/94.