Yes, IMHO it was a TERRIBLE decision. As was the recent one to go with IndexedDB instead of the far superior Gears-like Web SQL Database.
IndexedDB will likely take years to to match the performance of SQL+Sqlite, if ever, not to mention having to write dozens of lines of code to replace one SQL statement in the case of joins for instance. It appears to me to be targeted at lightweight use for SQL-less developers used to cookie-like put/get storage. It's a real stinker for complex database operations. This decision alone will set back browser-based development by years in regards to deploying really high-performance database-centric desktop-like apps. I almost never comment publicly but I felt so strongly about this issue I recently posted something criticizing the IndexedDB decision on the Mozilla forum discussing<http://hacks.mozilla.org/2010/06/beyond-html5-database-apis-and-the-road-to-indexeddb/>the matter and explaining exactly why I felt that technically it was such a poor decision. They deleted it and banned me from posting. While I should have been less harsh (I think I called the decision makers "Overeducated morons") clearly they are reacting poorly to the overwhelmingly negative comments. In fact it took me about 2 minutes to find a blog<http://paulrouget.com/e/lenovoshop>by the Mozilla moderator, Paul Rouget, who banned and insulted me, where he calls a website he used a "nightmare" and criticizes the designers. More than just a little of a double standard at work over there at Mozilla it seems, at least by some people. Unless I'm missing something they are replacing the working and extremely capable Web SQL Database with the new IndexedDB because it is more "aesthetic" to the web developers they discussed it with. "Aesthetic" apparently meaning to them you have to write many lines of object-like javascript to accomplish the same thing as one line of SQL currently. Even using their own examples they posted it clearly shows IndexDB takes considerably more code than Web SQL to accomplish the same thing, and it will get MUCH MUCH worse for complex database operations. In so doing they are not only throwing away a working, more capable implementation, but discarding the benefits of the years of work on the SQL query optimization rountines in Sqlite, because they are inserting their new "Aesthetic" IndexedDB layer in between where NO ONE will be able to write optimized complex queries. Perhaps I am missing something about their design and it is in fact superior to what we have now. I'm looking forward to someone explaining to me why it is better and not just a political decision essentially forced by Microsoft, who no doubt hates Web SQL and would love to see browser-based development of desktop-like apps set back a few years, if not forever, which is what I predict IndexedDB will do. What Google is thinking here is beyond me. Hopefully someone will come to their senses at some point and at least offer SQL pass through capabilities as they are all still using Sqlite on the backend, but who knows? On Mon, Mar 28, 2011 at 7:19 AM, <[email protected]> wrote: > That's the point. Gears is no longer supported by google. They are > phasing it out for html5 so soon when the new standard is made public then > gears will die. > > I don't think they should have done that as gears only really lived for a > few years. Just long enough for devs to build applications using the api > then once everyone was invested they cancelled it. > > So don't hold your breath for gears support in whatever app you mentioned. > Sorry. Hope it helps. > Sent from my Verizon Wireless BlackBerry > > -----Original Message----- > From: kmmontandon <[email protected]> > Sender: [email protected] > Date: Sun, 27 Mar 2011 23:03:28 > To: Gears Users<[email protected]> > Reply-To: [email protected] > Subject: [gears-users] Any plans on updating? > > Is Google Gears going to ever be updated, or is it pretty much dead? > > I'm asking because I'd like to be able to search my old Myspace > messages (yes, I know), and it's impossible, because Gears can no > longer search the Myspace PM database, since last fall. Seems like > this project is entirely defunct. >
