PC Week's Review of Warp 4

By Derek Clarkson

In the issue dated 30th October, PC Week did a review of Warp 4. For those of you who missed it, it had some very interesting things to say.

Essentially they didn't seem to think too much of it, in the section entitled "The Bottom Line" they said that "... the built-in Java support, will be useful for businesses wishing to deploy Java applications, but performance and integration of the Java Virtual Machine with Warp is sorely lacking." and that "The VoiceType navigation software is essentially unusable for mainstream PCs.".

However there was also comments like "Excellent support for PC Cards and power management makes Warp a better mobile client than Microsoft's Windows NT 4.0, which has limited support for PC Cards and power management."

An interesting chart was supplied which compares OS/2 Warp 4, NT Workstation 4 and WIndow'95, giving pros and cons of all three as well as a summary of features with how well they have been done. The interesting thing was that whilst their summary gives the impression that Warp is nothing special, when you look at the pros and cons you see quite a different thing. For example the Warp Pros stressed the MultiTasking, threading, Java, Network Admin tools, Scripting languages, etc. NT's stressed the '95 interface, remote access, networking, administration and multi hardware platforms. And in the cons, was Warp's slow Java machine which is not integrated well, installation 'still difficult' and the voicetype not working that well. Whereas NT's was 'Hefty' hardware requirements, lack of scripting, poor mobile support and licensing. It was not so much what was said as the impression gained from the whole.

Finally in this section was a chart listing different aspects of the operating systems and how they faired. Again Warp came out on top (Just!) with 6 A's, 13 B's and 2 C's. Whereas NT scored 6 A's, 12 B's, 2 C's and 1 D.

Finally in the review there was a section on Java and IBM's implementation of it in Warp 4. PC Week Labs did a series of performance tests using a benchmark from PC Magazine called JMark. This benchmark tests all aspects of the Java Virtual Machine across different options such as Warp's applet viewer, OS?2 Netscape Navigator 2.02, Javasoft's Java Virtual Machine in Java Development Kit, Microsoft Internet Explorer 3.0 and Netscape Navigator 3.0.

Overall, Warp's Java VM performed very poorly, being resoundly threashed by both the NT and Windows'95 implementations. In most test the Windows tests using Navigator and Explorer where at least twice as fast, on average probably 4 or 5 times faster. Only Javasoft's Java Virtual Machine was slower and then in only in a few tests, it was faster in just as many others.

This seemed rather surprising to me as I would have thought IBM would have achieved a faster rest, competitive if not beating the two Windows platforms. So Stay tuned as I intend to find out a bit more as to why these results have come up.