1 April 2005 | 0
I remember an incantation from the days when open systems were an even more hotly debated concept than today’s Linux versus The Forces of Darkness. It seemed then that the test of worthiness (let alone trustworthiness) for a computer system investment could be evaluated along three dimensions: scalability, portability and interoperability. There is some event driven need to perhaps include security in that short list of admirable system qualities. Then, as now, the flood of marketing puffery and unsubstantiated claims lead me to add a fourth test: credibility. All five need to be applied to proposed IT acquisitions to avoid well documented traps.
<p>You can scarcely turn around without bumping into a web services prophet these days. The calendar is awash with conferences, magazines replete with exhortations to jump aboard the bandwagon that will lift us all from the rather dull prospects of the winter of 2002. The IT industry is whistling in the dark as never before.
Even Gartner analyst Milind Govekar told a seminar in Dublin in February that web services were now: ‘At the very top of the hype cycle’. I can understand the sentiment but I suspect that we ain’t seen nothing yet. You see, Microsoft has just launched Visual Studio.Net (VS.N) the very thing needed to build the best web service future (according to Microsoft).
In this space I can’t lead you through a blow-by-blow evaluation of Microsoft’s latest bid to get everyone to build an Internet from which it can extract a toll every time you click. I reflect instead on those four immutable principles enunciated above, and, based on past performance, have qualms about the applicability of the fifth test, credibility.
The hail of brickbats has already begun regarding the security architecture of .net. We have all been touched by Chairman Bill’s words of contrition and his determination to halt the spread of features at the cost of security. The reality will always fall short of aspiration.
Microsoft has implemented a switch in Visual Studio.Net’s default security settings to prevent managed code from running from the Internet. This makes VS.N code safer to run on an intranet but blocks the hoped-for utility of making an application internet capable.
Once you have assessed the risks, you can switch the Framework’s Internet capabilities back on. Or at least that’s the theory. There have been a number of cases where users of the beta version of VS.N have been able to confuse Internet Explorer into thinking that the coast was clear when it in fact wasn’t.
Bugs will always be with us and one must applaud Microsoft’s late efforts to reduce the creepy crawly count in its products. But just because they now realise that there is a marketing value in appearing to be concerned, it doesn’t mean that their stuff will smell any better in the short term.
Just this week a buffer overflow flaw (not another!) was uncovered in executable code built by using Microsoft’s Visual C++.NET.
As for the other tests, scalability and portability have traditionally been a challenge for Microsoft. Interoperability seems to come along whenever Microsoft needs it.
We are well into the traditional battleground of de jure versus de facto standards. J2EE products from Sun, Borland, IBM and others seem to do better than Microsoft in terms of the first three tests. Security, the process, may be on the mend in all corners. Credibility I will leave to you to assess.
If web services are not just a hype-powered techno fashion then I supposed we have to keep our eyes open for ways that this great leap forward to transform our lives. To be sure, web services will be delivered with alacrity to the usual client suspects including the humble PC both in its tethered desktop and WiFi enabled notebook siblings. Thin-client computing (or Server Based Computing as Giga Group refers to it) is bound to experience a long deserved upsurge in popularity.
However there has been plenty of nudging and winking about a new generation of mobile clients that may be pressed into service for more ad hoc access to web services. Against this background the sudden outbreak in hostilities between Nokia and Microsoft is shown in sharp contrast.
We have already seen a glimmer of what is to come. The Handspring Treo, Motorola’s Accompli, the Nokia 9210 Communicator, the Carpe Diem I played with last summer and so forth. However, none of these really delivers on the promise of the web in your hands no matter where you are. Enter now a great disappointment called Pogo 1.
Pogo Technology has just thrust its immature mobile web (and more!) access device into the stream of commerce. Unless there is a pretty smartish upgrade then I predict a flop and a potential poisoning of the well for similar devices in the near future.
Check out www.pogo-tech.com and see just how compact and attractive a mobile web access device can be. The ergonomics are not bad and the list of features is quite comprehensive for a consumer-orientated device.
They include a dual band GSM phone, MP3 player, phone book, diary and a slew of games. The star of the show is a brilliant nearly 4in diagonal colour touch screen.
This screen is big enough for mature eyes to use in most lighting environments — a tough hurdle for smaller and less well-spec’d screens. The browser included is sufficiently compatible to render most web pages when they arrive including HTTPS and SSL security. Macromedia Flash is bundled and the execution of animations is done passably well. There is a MMC card slot for removable storage.
The blurb says ‘Who needs 3G?’ and there is more than a bit of truth to that rhetorical poser. However, the speeds in excess of a 56K landline modem promised were not delivered. Pogo 1 is said to be ‘GPRS ready’ but the network bundled with this UK£300 devices is not.
There in lies the first trip in Pogo 1’s downfall. It is as slow and flaky as WAP was back in the days that the phrase ‘Wap is crap’ was coined. If ever a device needed packet switching to show off its capabilities, this is the one. GPRS abounds in my bailiwick and it was a major screw up to not get aboard the ‘always on’ mode of data connection.
However, the second and third tumbles for Pogo 1 are fatal. One is the revenue model and the other is actually delivering the services that could justify my continued tolerance of such an expensive, ill-performing bit of techno frippery.
The tariff is standing-charge free (in the UK) and data calls are 10p (in this island of proprietary European currency) a minute. As you can imagine, you could run up a substantial bill very quickly especially with treacle-in-February responsiveness. On top of that, Pogo wants to mine your wealth for £8.99 a month for storing all your personal address book data on their server and supplying their early 2002 rendition of web services.
Those Pogo web services at the moment are extremely thin on the ground (aside from accessing the same web that you can from your deskbound device). Nonetheless, Pogo Tech is banging the drum of service providers with all its heart. Good luck to them on the chicken and egg nature of that quest. Subscribers are needed to attract services and services are needed to attract subscribers. Around and around we go!
I am taking a rather dim view at the moment of the corporate commitment to web delivered services. I know it will get better but for now it is dire. I tried to order my Pogo 1 from the Carphone Warehouse web site. It didn’t work so I ordered it by phone instead. When it arrived (bang on time as promised) I tried to register the phone on line. The site was down. The tech support hotline was not open the hours advertised on the company web site. Nor was the network provider’s.
I finally got it registered with the Fresh network (part of one2one) on the fourth attempt with much on-hold music under the bridge. I then tried to register the device on the Pogo service and that site was down and stayed down for at least a half day. After that, the tardy performance of the device and lack of interesting services drove the final nails in the coffin. I was cheerfully given a full refund five days after it was delivered. I can’t wait to see the air time bill!
Maybe it is a good thing that the tools needed to deliver a solid web service experience are only now being delivered. Commitment to delivering web based service 24×7 seems lacking.
When the hype dies down, companies will have a chance to reflect on the value of delivering services via the web and the risks that those services may not make it to the end of the UTP cable (or corner of the ether) required with concomitant loss of business.
We need web services only if they work better than what we have now. What is envisaged and what is ultimately possible may be two different things. The first four tests (scalability, portability, interoperability and security) need only be applied to emergent products if web services generically have credibility.