[LEAPSECS] ISO Influence

Poul-Henning Kamp phk at phk.freebsd.dk
Sat Dec 18 17:39:31 EST 2010

In message <89576419-9CA9-4FB3-83FF-05D958CE8ED7 at noao.edu>, Rob Seaman writes:

>I wrote:


>>> Then it should be straightforward to perform the inventory of

>>> the different segments, systems, and software to demonstrate one

>>> or the other of these assertions.


>Poul-Henning Kamp wrote:


>> How many lines of code can you review per hour ?


>The logic escapes me.

I'm just trying to put a lower bound on the budget of "straightforward"
in your claim above.

I assume you yourself are pretty good with source code, so if you
say you can review X lines of code per hour for possible leap-second
issue, we can use that to put a lower bound on the amount of time
it will take to flag all the pieces of code that may or may not
have to be worked on, depending on what the leap second decision
comes out to.

All you have to do is scan the source code and catch 100% of the
code that may have leap second trouble, and it's OK if you flag
a couple of percent that shouldn't have been, they will be
cheap to eliminate in the next round of review.

So assume you have the source files on your computer, and all
you have to produce is a list of sourcefiles with potential
for leap-second issues, and give a line number where you fear
there may be leap-second trouble. You don't have to explain
what the trouble might be, just the line number.

How many lines of code per hour can you do ?

> Due diligence in system engineering should not be controversial.

... and it should come with a cost estimate.


Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

More information about the LEAPSECS mailing list