It’s only been 3 months since I last reported in on ACT vendor assessments, and some changes we made to improve the match rate which you can expect to find when you sync and start looking for vendor data. Now that we’ve performed the update, I wanted to report back on how things have improved.

The good news is that the match rate on an ACT inventory is now more than double what it was back in April, when we began the latest initiative around improving that match rate. My hat goes off to the team who executed brilliantly against this objective – that’s going to end up saving a lot of time for people migrating away from Windows XP.

My friend Marc, who is a PM on the ACT team, went on to perform some additional analysis on match rate, to see if we can understand which factors contributed most to improving the success rate. Now, I had a hunch that simply being a company that operated outside of the US would be a factor, as anecdotally I found lower rates if I was working with a customer outside of the US, but I couldn’t prove that quantitatively. Unfortunately, we don’t have that attribute to data mine, and though he tried to approximate that by dividing those inventories that had at least one foreign language application in them, the outcome was utterly inconclusive.

What he did find, however, was a strong relationship between the match rate you can expect to find and the size of an inventory. Those with smaller, well managed application portfolios also had the side benefit of having a higher match rate! Here is how the data maps out (through August):

Hopefully, if you are using ACT, you have been able to benefit from the investments we made to improve the match rate and save yourself a bit of time. Perhaps now is also a good time to consider rationalizing away a number of your applications if your ACT inventory is bulging in the 100,000 range…

I had a question come up from a customer: so, IE9 JavaScript performance is faster, and IE9 has 4 different document modes – are there differences in JavaScript performance across the document modes?

Now, I kind of suspected that the answer would be yes, but I didn’t actually know any details – and as a general rule of thumb, I don’t like to just assume that what somebody else told me is true (or even that I am actually remembering correctly – I’m not getting any younger, you know) so I figured we would run a little experiment and see what happened.

I decided to use the SunSpider performance benchmark, which tests JavaScript performance. While there is a newer version available, that newer version doesn’t work with older rendering engines, so I had to run the older version. This is the specific test that I ran:

http://www.webkit.org/perf/sunspider-0.9/sunspider-driver.html

I then ran this test in each of the 4 rendering modes in IE9, and for good measure I also threw in a test of the latest Firefox and Chrome. These were the results of running the test in an uncontrolled environment, a single time, on my computer – I am not attempting to make a scientific statement about absolute performance, just to gain an understanding of whether there are differences which are large enough to matter.

Wow. The outcome seemed really clear to me: when in IE9 or IE8 standards mode, IE9 performance is very clearly competitive with other modern browsers. However, once you drop to the two oldest compatibility modes, performance drops by an order of magnitude – that’s a big deal!

Since Compatibility View is the default for the Local Intranet zone, I would therefore strongly consider a program of enforcing the inclusion of X-UA-Compatible headers to opt in to one of the modern document modes in order to maximize the performance of your enterprise web applications, which is precisely the advice I gave to my customer.

If your JavaScript performance is not what you’d hoped in IE9, perhaps you should press F12 and check your document mode…