The periodic question that many users of Microsoft Windows face is upon us again - to upgrade to the new Operating System (OS) or not?  This time it's Windows Vista, the long-anticipated successor to Windows XP.  For office computers, the choice is generally made by corporate IT groups, and it's usually a conservative choice of hanging back for a while.  For home computers, we're more at the mercy of the market.  The supply of XP loaded computers in the retail chain is drying up, so if you plan to buy a new PC in the near future, your only choice will be Vista.  But what about computers attached to laboratory instruments and automation?  Those PC's often fall into a "gray zone" of semi-support from corporate IT groups and in almost all cases they are in place primarily to run proprietary software for the control of or acquisition of data from an electronic laboratory device.  So the upgrade of existing laboratory computers to a new OS or whether newly purchased systems come with or require the new OS is completely dependent on the compatibility of that proprietary software with the new Operating System, and that of course is in the hands of the provider of that particular technology.

How does that process work?  To get some insights, The Lab Man talked to Dr. Bob Burger, Staff Software Development Engineer for Beckman Coulter, who is focused on their laboratory automation line of products, specifically their SAMI control software.  Bob has been with the company for 10 years, so he's had the opportunity to live through several of these transitions.  He indicates that how quickly they bring their software into compatibility with a new OS is driven somewhat by customer demand, the need to take advantage of any technical advances contained in the new OS and the potential lifetime of the new OS.  For instance, their SAMI product was originally written for Windows NT.  They subsequently skipped updating to Windows 2000, because WinXP was already on the horizon, there was no demand from their customer base and Win2k didn't offer any technical advantages for their product vs. WinNT.  They still have customers who are happily running NT4, and thus his software group still maintains the ability to run and test software on that version of Windows, even though Microsoft now no longer supports it.

According to Bob, that first revision of SAMI from WinNT to WinXP consumed several years of staff time, but the effort prepared them well for future OS revisions in two specific ways.  First, they rewrote SAMI in strict adherence to Microsoft XP developer guidelines.  This is supposed to make adaptations to future Microsoft changes easier, at least to the extent that Microsoft follows their own rules.  Secondly, they developed an automated test suite for the SAMI product.  This allows them to rigorously test their product against any new or revised OS with a minimum of effort and generate a report of any potential incompatibility for further investigation.  The Lab Man should point out that such a time investment in preparing for future life cycles of a product is much more likely to occur within more mature technology-providing companies and for products that have a foreseeable longevity.

Groups such as Bob's that are Microsoft Developer Network (MSDN) subscribers will typically get beta releases of a new OS several years before commercial release.  Those beta releases are quite changeable and are generally only useful in revealing broad directions or changes that might eventually appear.  The release to manufacturing (RTM), or final build of Vista was made available to MSDN subscribers late last fall, about 2 months before commercial release.  So you can see why it can be quite challenging to have a compatible upgrade version of a complex software package ready and available at the same time Microsoft commercially releases their new OS.

Bob indicates that through occasional early testing in the Vista beta days, followed by more intense testing with the Vista RTM version, they have successfully adapted their SAMI product to Vista.  However, they have not yet begun their own formal release testing process, a more rigorous testing procedure that would be triggered by a business decision to release a new version of the product.  Many laboratory automation users may have the misimpression that companies of the size of Beckman Coulter have literally fleets of software developers to throw at such problems, but in reality these project teams are relatively small - usually 3 to 4 people.  When they do develop new software, such as a Vista upgrade, Bob says that one of their key practices as developers is to "eat their own dog food"!  In other words, they will use the software they develop and use Vista to assure it "tastes good" before releasing it.

Listen for sounds of contented munching if you happen to pass through Indianapolis!

As always, your comments are welcome and encouraged!  Listen to the Podcasts!!

Domo Arigato, Mr. Roboto