There is a lot of raging back and forth about this on the net. Here is
another take on what was said by the blogger you mention above.
"They are not global temperatures; the error only applies to the US data, and the corrections apply only to the US data. Unless the bug is bigger than either Steve McIntyre or GISS have disclosed, 1998 is still the officially the hottest year on record for the global temperature, and 1934 is only officially hotter in the US. Ace, you're getting a little carried away here. That being said, I would characterize the US network as being the most extensive and most studied; god only knows how unreliable some of the non-US data is that goes into the global record. That being said, the real story is not a few little corrections here and there by around 0.2 degrees C. The real story is that GISS data was obviously wrong but had gone unchecked before Steve's work,
firstly because the IPCC choirboys never bother to check their own data (see also the MBH98 hockey stick), and secondly because it was hard to check because the GISS software and algorithms are kept secret (despite being paid for by public money). This makes it virtually impossible to replicate and check. Replication is a cornerstone of the scientific method; its not just bad science, its not really even science at all."
And I would add that we humans seem more interested in the battle for political control and the delight in being
right than in whether we are going to face a possible end to life as we know it on the earth. That, it seems to me, is the real story.