Netscape Communications Corp has discovered a new religion – crossware, applications for the networked enterprise that will be a new breed of ‘ubiquitously available, service-based,’ ‘intuitive and easy to use,’ ‘scalable to millions [of users],’ ‘low-cost [in terms of] development and maintenance,’ custom applications that can be easily extended to external partners and customers. Its chief technology officer Marc Andreessen used the first day keynote slot of the company’s developer conference in San Jose, California as a platform to lead the singing of the new hosannas from on high.
By Gary Flood
The day previously, Netscape announced a set of initiatives which not so much mark a new direction as simply reinforce existing statements of direction, though all furthers the crusade. Firstly, Netscape has adopted Sun Microsystems Inc’s JavaBeans component model for its upcoming Netscape One – Open Network Environment – platform, and the two, in concert with anti- Microsoft buddies IBM Corp and Oracle Corp (CI No 3,118), are to bring JavaBeans and the Object Management Group’s Corba Common Object Request Broker Architecture standard closer together, by outlining what they think needs to be done to Corba for it to be able to support JavaBeans. Andreessen has outlined the road map Netscape envisages for achievement of crossware. This will include full integration of JavaBeans in future Netscape products; support for multiple, non-Java, language support for JavaBean development in Netscape including HTML and C/C++; a technology called BeanConnect for connecting multiple JavaBeans on the same HTML page to build Java applications; a crossware directory of components – Netscape claims 90 such functions are already available; a Visual JavaScript tool for JavaBean assembly; and a new JavaScript debugging tool written entirely in Java.
New religion
What’s wrong with this new religion? Are we being so churlish as to suggest that Java is not Most Righteous and that MSFT’s ActiveX is not yeah verily the Devil’s Spawn in contrast? Not at all. What worries us about this eschatology is that a very new language called Java – with JavaBeans-Corba extensions, to be sure – and a web browser will at a stroke provide the application development needs that will satisfy any and all enterprises and at a stroke solve all of the information service’s department’s problems of the last forty years of data processing. Ah, if only by truly believing we could make this particular Tinkerbell come back! Yes, over 100 leading companies may well have chosen Netscape’s new client – Communicator and server – SuiteSpot 3.0 as bases for either corporate messaging solutions or for implementing ‘business critical’ applications for their ‘intranets and extranets,’ as Netscape is crowing. But in the very same press release we get told that such mission-critical applications include better calendaring, messaging, and web navigation; making previously only internally-available publications externally available on the web; giving users better access to their medical and dental benefit information; and as a replacement, standardized electronic-mail system. Pardon our ingratitude to the Prophets of Java, but neat as those systems are they probably don’t turn the power on in the morning or make sure the checks get cut every month. And you can do lots of great things with a browser, or even a network computer, but shucks, what about when you want to do stuff on the server? What application development metaphor will we use then – a coffee- based one or a tool-based one? Watching the demonstration of how easy it is to build and extend a crossware application the heretical thought that crossed this reporter’s mind was ‘Steve Jobs used to do this demo better, with Nextstep, six years ago’; and that this sounded awfully like the Millennial pronouncements of the Object shamen at the start of the decade, only we say ‘components’ now instead of ‘objects.’ Rene Descartes used kind of the sneaky ‘ontological’ argument proving the existence of G
od, which basically suggested that just being able to imagine such a Great Being proved he must Exist. Crossware (and rather sweetly Netscape makes a special point in its white paper of giving credit where it’s due for this revolutionary concept – so huzzahs for Brian F Kathman, an internet/intranet consultant with Downer’s Grove Illinois-based Systems Resources Consulting!) is using an ontological ‘proof’ for the existence of crossware; it sounds as if the technology should exist, so it must.
Theological argument
Using this approach may well result in portable applications, but is Andreessen pretending that the vast bulk of application development at the desktop level is not being done in Microsoft’s Visual Basic these days – so which applications will prove to be portable? Yes, the Java Development Kit 1.1 may well have over a million downloads to its credit, but using that as a theological argument to suggest all custom application development going forward is to use Java is just plain daftness worthy of a counter-reformation. Crossware that embraces ActiveX wouldn’t even be crossware in the sense Andreessen suggests either, because it still just means Java-friendly applications that use Netscape as the central palette, and do we all believe we can launch all the apps we will ever need from the browser? Another failed crossware-like technology’s ghost was mentioned at the press conference held prior to the keynote – OpenDoc, which only last year IBM embraced as another sort of cross-platform holy grail. And do we even need to bring up legacy code, which Marc may not want to think about because it’s not very cool and you probably didn’t wear a polo shirt while writing it, but the relevance of same he may have heard about if he’s ever read about this Year 2000 malarkey? Crossware is a great concept and we truly wish this work by Netscape et al had achieved it, but what it actually is a philosophical abstraction which would be nice to have. Let’s all stop speaking in tongues about what Java can do and will do, and maybe – just maybe – we might get nearer to actually making something like it really happen.
