@Kizule: A bit late (since only today I saw your patches to this extension), but my $0.02 on this...
LinkSuggest (the original version) was written by Inez Korczyński at Wikia either in 2008 or in 2007; it used YAHOO! User Interface (YUI) JS library, since Wikia used that at the time. Eventually Wikia transitioned to using jQuery and mainline MediaWiki releases started including jQuery and the jQuery UI library as well.
In late March 2011 I committed a fork of Wikia's LinkSuggest under the same name to the WMF SVN; the vast majority of the meaningful JS fixes weren't done by me, but rather by the brilliant Ciencia Al Poder. He rewrote the extension's main JS part to use the jQuery UI autocomplete widget instead of YUI, which made the extension actually usable since you didn't need to include another heavy JS library with it to use it, but instead you could just use the JS library that shipped with MediaWiki. (And, curiously enough, at least in August 2011 a Wikia developer contributed code to the upstream version of LinkSuggest, suggesting that Wikia may have used my and Ciencia's fork on their sites at that point instead of the original version they had developed, which had the YUI dependency.)
That's basically it. Ever since late March 2011 LinkSuggest has been maintained and functional.
This extension was committed to the WMF gerrit in mid-February 2015, about four years after the LinkSuggest extension by me and Ciencia became available. Now, this version has some interesting approaches I'm sure would have been lovely to incorporate in the original LinkSuggest, though it should be noted that LinkSuggest2 is GPLv3 only whereas the original one is GPL 2.0 or newer.
Forking is a fundamental right associated with open source software and it's important that such an ability exists, although I'm not sure how much of LinkSuggest2 is really forked from the original version, besides i18n. That said, it's usually way more productive to collaborate together. Speaking from years of experience (of maintaining other people's awful intentional design decisions and the like), I know I'm biased towards avoiding large refactorings and the like, though MediaWiki core these days seems to be to the contrary. That said, I don't recall there being any discussions whatsoever prior to LinkSuggest2 becoming a thing. I'm definitely not opposed to having such talks now and discuss on how we can best improve the software and create the ultimate, definitive version of LinkSuggest instead of having multiple different approaches for the same end result. After all, volunteers can only maintain so many things effectively, so having efforts split is kinda counterproductive.