Git vs. Mercurial analysis

I'm only a couple paragraphs in, but there's already some pretty heavy bias here that's turning me off from continuing.

First off, the recommendation right off the bat is (paraphrased) "Companies should use Mercurial because it's easier to learn, which translates to saving time and money". Even if I accept that it's easier to learn, this completely ignores any consideration about interoperability with existing open source projects that you may or may not need, or the existence of any tools built around git or mercurial that you may want to use. The learning curve is but one factor that should be taken into consideration; it is not the only factor.

Secondly, the very first actual comparison between the two is highly disingenuous. It makes a claim about the documentation for `git commit` vs `hg commit` using what it says the summary of each command is. But that's not the summary. That's just the first line of the in-depth documentation. The actual summary of both commands, which is the very first real line of text you see in the help on both, is as follows:

    git-commit - Record changes to the repository

pretty straightforward. And for `hg commit`:

    commit the specified files or all outstanding changes

Also pretty straightforward. But these lines have been completely bypassed in order to provide a comparison that uses more technical language on the part of git and therefore is perhaps less understandable for a nontechnical user.

---

If we continue onwards, we find more absurd comparisons. There's a nice little ding against git here for actually listing its options in the synopsis, which is standard behavior for unix utilities, unlike hg which just says `[OPTION]...`. Sorry Robert, but git definitely wins that one.

And then he goes on to complain about terminology used in the description of `git checkout`, such as "working tree", even though these terms are well defined for git and are used consistently in all documentation (definitely a plus). But he doesn't mind terminology used in `hg update`, such as "working directory" or "changeset". You can't do an objective comparison between two VCS's learning curve unless you're willing to actually assume the user has no prior knowledge of the VCS. Big fail on Robert's part. And it's enough to make me stop reading here.

This isn't an analysis of Git vs. Mercurial, this is an analysis of Git's documentation vs Mercurial's documentation. Disappointing.

Just a quick note -- if you don't want to make an exception for the certificate, you can go to the page over http.