Freshmeat used to be a milestone of opensource software for a long time. It offered a good service for developers, including APIs, to publish your software and have it categorized. You hit the frontpage for every version, which would be a very nice publicity boost even for scarcely known applications. If you search through Ubuntu/Debian's repository you can still find a lot of helpers to "release" on freshmeat, which should give you an idea of how much freshmeat was entrenched in FOSS.
As an user, freshmeat was great. You could search by category, release type, language, environment, etc. Pretty much everything that was afterwards integrated in the sourceforge "trove" was pioneered by freshmeat. It was my go-to source for software just after an "apt search". Yeah, the look was funky, but who cares? Most of the software was so technical that many simply didn't care.
Google would (and still does) return too much crap for pretty much any query regarding software, whereas freshmeat offered immediate insight though filtering, categorization and project statistics. I've found so many gems on freshmeat that I would have never have found in Google (because of the sheer amount of results - mine being always one of the last ones), stackoverflow (field is too narrow) or github (which basically restricts to software developed with git, and is often incomplete).
They decided to redesign it a couple of years ago, and while they were at it they decided to throw away anything that was still good about it: search. Of course, the new trend was tag clouds, so you can guess, now projects are now categorized by tags. As such, it's now pretty much impossible to guess the correct tag/tags for a project. It's either a tag with a different spelling or a different word altogether. The tagging process is also too liberal: in the old freshmeat, you were encouraged to categorize your software in several hierarchical groups. As such, many projects would have a proper "programming language" category filled it, a proper "operating environment" (like command line, X11), etc. Searching was a breeze just because you were guided during the categorization.
Now it's basically the same as a google search: you either have too many results, or zero. Giving the programmers just a "tags" field to fill it has resulted in the archive having absolutely no consistency.
It's a shame, really, because there's a lot of niche software (visualization, scientific tools operating of weird data, a lot of old stuff) that I can still only find on freshmeat.
I've been thinking myself of re-implementing the old site simply as a rip-off just to have this archive back. It's sad that all they would think of now is change the name, because that's irrelevant for the highly technical users which used to publish/search on freshmeat.