Image of Navigational Map to Home / Contents / Search ToyBox

by Peter Wone - GUI Computing

Image of Line break

My mission this year - apart from a couple of things I'm not allowed to discuss - is basically to thoroughly investigate web related technologies.

Where will the future lie? Do we bet, as always, on Microsoft? Or has good old uncle Bill finally missed the bus, which is another way to ask does Netscape really have web technology under its thumb?

Netscape's technologies are fundamentally simpler in approach. This is partly because a cornerstone of their strategy has always been true platform independence, which keeps them from getting architecture confused with implementation.

This is a choice. Microsoft, on the other hand, went the other way. That's why their stuff always appears first and often only for Intel boxes, and if it ever gets ported, it's obviously a port of something built for another platform.

But the world is filled with Intel boxes. I've got one. I'm using it now, and it's a very good bet that the same is true for you, so it's not such a bad thing for Microsoft to make that assumption. Certainly it permits liberties not available to Netscape.

Let's have a little retrospection here. Probably a number of you have at one time or another played with Microsoft's own little Internet, the Microsoft Network (MSN).

MSN is managed. You can do a net-wide search using a nice tidy little built-in find dialog much like the one Win95 provides for searching the local LAN. It's fast and simple and easy to use, and unlike searches on the Internet, it is exhaustive.

But it's big and slow and it's proprietary software, which means that if Microsoft doesn't feel like fixing something then nobody will. Their viewers are slow, and the caching strategy is so slow that I suspect there isn't one.

The MSN client (browser) is a no-brainer to use, but the old adage holds and it seems to attract people who can most tactfully be described as inexperienced or lacking in netiquette or both. That would be OK, but it's sufficiently easy to use that there is no inducement to learn. For the first time ever I am beginning to wonder if there is such a thing as too easy to use.

You can't really extend MSN. The Microsoft people are about to tell me that you can, but you have to get the SDKs from Microsoft and there are still limits to the scope of your creativity. And being a service provider on MSN requires you to buy into a bunch of obligations. There's a contract.

The Internet, on the other hand, isn't a managed thing. There's some serious free-enterprise happening here. You live or die on the strength of your server, your applicability to current market (user) needs and your talent at second-guessing next year. You can quickly create new markets, if you have vision and venture capital.

And venture capital requirements are orders of magnitude lower than for manufacturing. Any programming house already has the basic resources, provided it can afford to take them off-line to gamble that the market will be interested in the vision.

MSN itself is irrelevant. While it's not actually dead yet, Microsoft has redirected its efforts to internet technologies instead, which amounts to an admission to the points I just made. But there are still issues to consider.

We have two strategies in front of us.

Netscape offers us platform and vendor independence in the form of Netscape plug-ins, Java and JavaScript.

It's not totally unstructured - the plug-ins follow Netscape specifications, and Sun and the UNIX community dictate Java syntax. Java is interpreted, which implies platform independence - provided that an interpreter is available on the target box. JavaScript is basically tiny Java with the necessary interpreter built into the Netscape client software (browser).

Microsoft offers limited vendor independence. ActiveX components are an Internet implementation of the idea embodied by VBX and OCX controls for Visual Basic and other Microsoft languages like Visual C++ 4.0. When you try to view a page which makes use of ActiveX technology, your client software will check the registry for the presence of the required version of the ActiveX control, in much the same way that any web browser checks for the presence in the cache of graphics. If they're not present, they get downloaded.

Much has been made of security issues. Automatic download and run of executables does indeed pose a serious security risk.

Sun's solution is to knobble Java so that it can't write to the local disk, and JavaScript deliberately has the same limitations. Netscape touts this as a very safe approach, but it isn't. As an exercise, a friend and I set out to find ways around it, and we succeeded via the Netscape Helper-App interface. Helper-Apps can be disabled, but this eliminates most of the useful possibilities for the browser and would render (pun) many sites unusable.

Since the Netscape implementation of the Sun security strategy is insecure, digital signatures are essential. That leads us back to the Microsoft strategy. It also makes it harder (but not impossible) for the little bloke to participate.

I draw the analogy to OCX technology quite deliberately. The two bear more than a passing resemblance, in terms of both structure and implementation, but that doesn't guarantee ActiveX the same kind of success that OCX technology can expect. On the desktop Microsoft has a stranglehold because it owns the infrastructure. Internet infrastructure belongs either to nobody or to the UNIX community, depending on your point of view, and Microsoft must suggest - rather than dictate - software architecture.

The next paragraph is not suitable for people of limited patience and may frustrate some readers.

At this point it's customary to round off ToyBox with a conclusion full of (hopefully) profound insight, but not this time. This is the most volatile topic imaginable in the most volatile industry imaginable. I don't yet know what will happen next, so I'll just waffle on for a bit longer, pointing out some more salient facts, and leave you to draw your own conclusions. ("The consequences of this situation are left as an exercise for the reader.")

It's very possible that the world will split into two camps, as it did with the Macintosh. This time, however, there's a difference. The Netscape school of thought utterly dominates the UNIX world. Not only is the UNIX world big and ugly enough to successfully ignore Microsoft, but it habitually does so.

Add to this the fact that all the UNIX tools are (a) two or three generations more mature than their Microsoft counterparts and (b) inherently faster for technical reasons too involved to discuss here, and you will see that unlike the PC/Mac schism, this is an issue that won't go away on its own.

Written by: Peter Wone
April 96

Image of arrow to previous article
Image of Line break