Habrahabr not a cake. Habrahabr cheese.

To our common regret, the Habrakhabr autumn update has many convincing signs of a raw code (read them and add them in the comments if I haven’t noticed or missed something ):

• Designation of several HTML elements in comments to blog posts fell off: blockquote, source, em, strong, sub, sup, etc. It is interesting that some of the elements work uninterruptedly when the user views the list of his comments on the page habrahabr.ru/users/ username / comments /
  
• The update button and especially the comment jump button became narrower in height — and the difficulty of hitting it, according to Fitts’s law , increased dramatically (the graph of the logarithmic function becomes steeper when it approaches the ordinate axis on the right in the direction opposite to the x-axis).
  
• The info class div elements have bitten off the right inner field, so that the check marks for voting for a comment are not aesthetically close to the right edge of it.
  
• In the blog posts, one chopped font (Verdana) is used, and in the comments it is slightly different (Arial), the contrast between which is too insufficient, therefore, unfortunately, it does not cause a feeling of conscious difference, but of sloppy under-unification. I’ll say right away that actually the transition from Verdana to Arial is rather justified, because, as you know , the original version of Verdana, which has been supplied by Microsoft since 1996 as part of its entire product line (Microsoft Windows, Microsoft Office, Internet Explorer for Windows, Internet Explorer for Mac OS ...), contradicted the Unicode standard, since the characters of combining diacritics (accents, umlauts, etc.) were superimposed not on the previous character, but on the next one. This bug has been fixed only in the version of Verdana font that is included with Windows Vista and subsequent products, and also comes as part of European Union Expansion Font Update ; but they are installed, of course, not at all, so the use of Verdana should generally be avoided on the Internet in order not to create a disparity with accents. But the current transition to Arial is not complete, and therefore half in, so it looks damp, and that's his problem.
  
• The elimination of the “Everything” category (in the meaning of “All records”) also looks raw - that's just not that it is not sufficiently decorated, but rather not well thought out. It turned out that there is no way to read new blog posts on company blogs without having to go and love (by clicking the button) each company separately - so that it appears in the ribbon settings. And the emergence of new companies and blogs in general, one must think, will pass unnoticed, because, in order to notice them, one must read them, and in order to read them, one must first notice them and go to check the options in the ribbon. Vicious vicious circle. But Habrahabr exists on the money of the companies leading their blogs on it, and therefore should be interested in attracting new paid bloggers - so the autumn update in 2011 looks frighteningly close to financial suicide. I would not like to see Habrahabr following in the footsteps of the unfortunate Avianova.
  
• In some parts of the site (for example, on the page where the blog text is typed) the javascript library mootools is used, and in others (for example, in the blog recording itself) jQuery is used. I’ll say right away that in general, the transition from mootools to jQuery is more likely to be justified, but it is clear that it is not complete and that some of the behavior on jQuery could not be transferred (or could, but still far from the way it was before - say, an order jump comments somehow changed chronological).

Do you know what suspicions this is all about?

Suspicions are that when developing Habrahabr DVCS is not used (for example, Git or Mercurial) with simple code branching as new functions are progressively introduced, but simple VCS is used (for example, SVN or even CVS), so that some changes in the site code are generally it is impossible to separate from others (because all changes, even raw ones, are progressively made in the same branch of code, in the same pile of files). And it turns out that when the site administration demanded to immediately introduce just one new product (for example, cancellation of invites), then, of course, it had to implement a whole bunch of other raw and unsecured code with two libraries instead of one, with two fonts instead of one, with clumsy CSS and fallen off source code analyzer in the <source> ... </ source> element.
')
You can hardly think of a better example (and even advertising, advertising!) In favor of Git or Mercurial, and in general in favor of DVCS systems, than the current circumstances of Habrahabr.

That is why I originally posted this blog post on the Version Management System blog — and not at all on the I Habrahabr blog , as one might think.