It has to be quite limited. Otherwise for example, lets suppose I have an integer (encrypted of course) and I have comparison and addition/subtraction and multiply/divide.

I can very easily find the encrypted values of both 0 (a-a for any a) and 1 (a/a)

I can now decrypt the data with repeated additions (or subtractions) of 1 and equality comparisons.

And, I don't see how you can prevent equality tests in the encrypted domain. You might have to calculate a Kernel but surely there is no way to prevent that.

So I don't see how the operations available can be as much as the usual operators on reals.

If they are responsible as well for implementing or administrating the system generating the tickets, then arrival rates need to also be measured. You don't want the team creating easy to fix outages so that they can bias the metrics.

In general you need to try to work out what behaviors a given metric will tend to, or could, produce, and you need to combine the metric with elements that measure unwanted behaviors. You want the resulting metrics to be scale free, so that the metric cannot be gamed simply by changing some parameter.

Simple example, suppsoe you measure a maintenance team by how many new regressions they create (per month say) in the maintained system. The team can get zero (which in this case is good) by never fixing any existing bugs). So as a metric this is useless. Instead of course, you should be measuring regressions per bug fixed. This is scale free because you can measure over one month or one year and the size of the values will not change just because the period changed.

So this sounds like an urban myth to me.

The problem is performance creep. Its easy to find big slowdowns with regression analysis, but you get perhaps as much a 5% variations in timings just depending on the phase of the moon. So any slowdown less than, say, 5% is not discernable from noise. As a result your performance can deteriorate by a few percent per checkin. Over a year that can mount up.

So you have to combat this by actually tuning the software every so often - say once per release - to recover the creep. And, of course, after you do this a couple of times, it gets harder to knock hot spots on the head and you have to do it early in the release cycle as you have to start rearchitecting to actually make a difference.

I wonder if you can still get thermal paper for a silent 700?

That's an editing picture. My COBOL is very rusty but I don't think you can pull just the digits out of it:

        77 SOCIAL-SECURITY-NUMBER PIC 999-99-9999.
        77 NUMERIC-SOCIAL PIC 999999999 COMP-3. ....

          MOVE SOCIAL-SECURITY-NUMBER TO NUMERIC-SOCIAL.

IIRC, this does not work.

(the patent app is still crap though)

When they want to send a message, they hatch an egg, and then tie the message to one of the chicken's legs before sending it flying back to base (I forgot to say they are also issued a catapult)

Probably no less plausible an approach than these "infofuses".

I soak the belt in water before I set off and then I can always find north by checking on which side the moss is growing.

Join the Army! Travel the World! Meet lots of interesting people, and kill them. (Spike Milligan; The Greate McGonigal, the Truth at Last)

So why does every update to Windows cost me $100?