Technobabbles I try to sound like I know what I'm talking about. Don't be fooled.

16Jan/090

Comments On Google Changes

Sign at the GoogleplexImage via WikipediaOn Wednesday, Google announced changes to or shutdowns of several services. Google also shut down Lively, the 3D chatroom that I always looked upon as a rather silly Second Life knock-off, beginning January 1 — just a few short weeks ago. ("Reasons Google should kill Lively" was actually a topic in my to-blog list for several months, but it looks like I didn't have to blog about it for Google to see that it wasn't a good fit with their other projects. Nice work saving effort there, self. :P Anyway...)

I'm rather unconcerned with the fate of Google Catalog Search, which was (I believe) really just a good way to work on the OCR technology Google now uses in Book Search. (Catalog Search's former homepage at catalogs.google.com now redirects to Google's main site.) I never used it; catalogs are pretty useless these days anyway what with online shopping and Froogle (now known by the much-less-punny appellation of Google Product Search and accessed by a link in the Google header called "Shopping", though I wish they'd bring back the old name).

Image representing Google Notebook as depicted...Image via CrunchBaseAlso of little real consequence to me, personally, is the development stoppage on Google Notebook. I don't use it much, and my service will be unaffected anyway. Well, relatively unaffected, at any rate. When I do use Notebook, it's usually in conjunction with the "Clip" function of the accompanying Firefox extension — which will no longer work. But the service will continue for now as long as one already has an account.

I'm actually somewhat glad to hear of Dodgeball.com closing and Jaiku being open-sourced. Dodgeball was a premature service somewhat like Brightkite — which I occasionally use — that stagnated almost immediately after being acquired by Google in 2005. Its interface has always been phone-only (Brightkite allows use via text message, Web interface, or iPhone/iPod-Touch – optimized site). In my opinion, Google would do well to encourage Dodgeball's users to move to Brightkite. An agreement with Brightkite to ease the transition for users willing to make the switch would likely make Dodgeball's death as swift and painless as possible.

Image representing Jaiku as depicted in CrunchBaseImage via CrunchBaseJaiku has also stagnated, the most notable annoyance being that it has been invite-only since its acquisition in 2007. With the transition to open-source (the service will continue to be run by a team of volunteer Googlers), Laconica1 might get some new features, competition, or perhaps both. Twitter may also be encouraged to develop long-overdue features like OAuth support, since Jaiku is slated to support OAuth right out of the gate when it is released to the open-source community. (Securely logging into Twitter from third-party applications and websites has long been a point of contention in the community, because the only option continues to be giving every app your username and password. Not a very secure solution, especially because there isn't even the layer of security provided by API keys such as used by FriendFeed and, yes, Jaiku. Google itself has supported OAuth authentication for its own services since as long as I can remember.)

Overall I think Jaiku's fate is the best and most (potentially) beneficial of all those announced this week. The potential for competition and improvement in the entire microblogging and status-update ecosystem is wonderful. However, potential users of the newly – open-sourced Jaiku Engine (hopefully that's the address where it will live — it 404s at the moment, but I'll watch and update if necessary) will still be dependent upon other Google amenities, namely Google App Engine. Jaiku was ported to App Engine last year, and the code base is now dependent upon being run in that environment. (This could be a ploy to get money, since App Engine charges — or will charge; I'm not sure of the time frame of the fee structure—nominal fees for applications that move beyond moderate-scale deployment — not a bad business move, if I do say so myself.)

Google Video PlayerImage via WikipediaThe one impending closure that I am really and truly saddened by is the shutdown of uploads to Google Video, which is now (and will shortly be, but for already-uploaded content) solely a meta – video-search site. I've always preferred Google Video to YouTube for a variety of reasons. Some of them are:

  • cleaner interface
  • more professional player appearance (nice for embedding on sites like swrobotics.com)
    • especially the lack of related videos and pop-down search bar
  • less cluttered site
  • fewer extraneous features
  • more focus on videos, less on social networking

(As an aside, there is also some uncertainty how the closure will affect users of Blogger who use the platform's video upload feature, which uses Google Video for hosting. I have never uploaded a video for a post through Blogger, so I'm somewhat detached from this particular concern — but I thought it was relevant nonetheless.)

I have many more reasons and thoughts on this comparison that I can't easily articulate, but I will be very sad to lose the ability to add new content to Google Video. Failing the motivation to deal with YouTube (which has limits on content length and filesize that may or may not be lifted in the aftermath of Video's shutdown), I suppose I might have to start uploading somewhere else entirely, like Vimeo (whose player I like quite a bit).

YouTube has always kind of irked me as a place to upload content. It's a great place to go to watch videos, almost always, but for hosting videos intended for display on another site... Despite the number of sites that do so, its player has always seemed out of place on the sites I'm involved with. The in-built social network (which includes profile pages, a messaging feature, "friendship", and so forth) has always seemed like an unnecessary layer to me.

I suppose I should have seen the writing on the wall when Google's own blogs, which used Google Video uploads themselves for a while even after the YouTube acquisition, all switched over to YouTube videos and left Google Video in the dust. Maybe I'm a video Scrooge, or maybe I'm just being resistant to change (who isn't?). Whatever the reason, I — honest and truly — will miss being able to add my videos to such a simplistically elegant site as Google Video has been.

Update (19:46): Some good news (depending on how you look at it) came this week, too: FeedBurner now shows a link to migrate your account to Google on the My Feeds page. That wasn't there a couple days ago...

  1. Update (2010-05-04): Dead link removed; project was renamed to StatusNet a while ago. []

13May/080

Google Friend Connect: Interesting Service, Can’t Sign Up

Yesterday, Google launched a preview release of Google Friend Connect at Campfire One. The premise is neat: Google provides you with a directory of pre-programmed social widgets (slash gadgets, slash modules, slash whatever you want to call them) that you can install on your website simply by pasting HTML code. All kinds of authentication formats are accepted, from Google Accounts to OpenID and beyond. For more nitty-gritty details, head for the official blog post or the Google Webmaster Central version (the latter is simpler).

Incidentally, there was a huge blogstorm about this yesterday, before it was officially announced and before the site was available. Rex Hammock thinks it might be the highest single-day pageview count for a 404 Error page, as he briefly states in a post entitled "Google's new Not Found feature is really fun". Go ahead, it's worth a click. (Well, Ctrl+click or Shift+click if you want to keep reading this ;-) but you already knew that, right?)

Sign-Up Error?

So, it sounds interesting enough that I want to get in on it. Like Google App Engine (my own coverage) from last month, there's a waiting list. (I got into App Engine, by the way, and immediately wasted one of my three indelible applications with a useless test listing I'll never code. There really should be a delete function in App Engine, but I digress...)

Unlike App Engine, though, this waiting list is powered by a Google Spreadsheets form. And this one doesn't work. I get a nice generic error message, which leaves me totally in the dark as to what went wrong. Argh. (It did get fixed though; see update below.)

Until it's working (I'll speculate why it's borked in just a moment), I'll paste my additional info paragraph here, just as an example (who knows, you might want to use it as a template :D ):

Initially I plan to just experiment with the service and see what it can do, but I think I might come up with a brilliant idea to actually use it for real as part of one of the sites I manage. http://www.swrobotics.com/ comes to mind as another possible site where I'd use it.

So anyway, why might this Google sign-up form not be working?

Since it's based on a Google Spreadsheet, it's bound by the limitations of the platform. Spreadsheets are limited (only one thread of many) to 10,000 rows, and with the publicity this launch got it's completely believable that 10,000 developers already applied. This is the price for Google eating its own dog food (so to speak). I guess in a way it's a good thing that they don't give themselves superpowers in their own services.

Meanwhile, I can't sign up for the waiting list. Oh well, it's not like I'd get immediate access anyway. Being locked out has increased my interest, too, so perhaps using a limited submission system was a calculated move on Google's part to generate even more desire to sign up. Who knows? It's working on me.

Anyone who gets in on it, please let me know what the interface is like from the inside. I'll probably post again if and when I gain access myself, but advance knowledge would be nice to have. (Yes, I will be watching Google blogs, official and unofficial, for more info as well.)

Use Cases

Meanwhile, I can brainstorm. This site already has social features (comments), but perhaps I'd like to do more. A sidebar game or something. Perhaps not, but you get the idea.

The Southwest Robotics website might have a comments area added where people can chat about the team's efforts or robots in general. There's definite potential in that idea. I'll have to take it up with my teammates.

Other than those few examples, I'm sure there a vast number of gadgets that I don't know about yet. Only getting into the service will allow me to really see what's possible. Stay tuned for more opinions as I can generate them.

Update (05/14): OK, the form is working again. It's also been completely changed; the number of fields has at least doubled. So much for just copying and pasting what I tried to send in yesterday.

[Google Friend Connect logo from the Google site, displayed under a Fair Use claim; error screenshot taken myself, site design and imagery copyright Google, also Fair Use claim]

8Apr/080

Google App Engine Launches

Apparently, TechCrunch's speculation that Google would be launching BigTable as a Web service was well-founded. Last night, BigTable -- along with a bunch of other amenities -- went live as part of Google App Engine, a service that could potentially compete with Amazon Web Services' EC2, S3, and SimpleDB offerings.

What happened last night was a preview launch, open to the first 10,000 developers who signed up for the service. By 09:30 (UTC), all the open slots had been taken, relegating me to the waiting list. It's perfectly fine, because I couldn't even start to play around with the service yet, much less use it for something practical.

The problem there is the fact that the service only supports Python as the development language right now. Google states (through the use of the word "currently" in the relevant FAQ entry and more explicitly in the documentation) that it's only a temporary limitation and that they hope to support other languages in the future, but it means that I have to learn a whole new language if I want to play around with the service right now. (Not that I can until I get my invitation...)

The Google App Engine features are as follows:

  • dynamic web serving, with full support for common web technologies
  • persistent storage with queries, sorting and transactions
  • automatic scaling and load balancing
  • APIs for authenticating users and sending email using Google Accounts
  • a fully featured local development environment that simulates Google App Engine on your computer

Contrary to what some people have thought, users of applications built on this platform do not have to use Google Account logins; it's just another API available from the environment, and doesn't preclude implementing a home-grown authentication system.

Google says the service should be able to absorb large traffic spikes with no trouble whatsoever. I believe it, if only because this platform is basically the same suite of tools Google uses to run its own applications.

One thing that bugs me -- besides the Python-only issue -- is the fact that Google App Engine uses a query language called GQL (I'm guessing the G stands for "Google") rather than industry-standard SQL. I suppose it's OK, though, because strictly speaking, BigTable is not a relational database. SQL-style JOIN queries are not supported, and GQL can be used with multiple types of data (it's used for the entire data store within App Engine).

So what are the limitations on the service as it's currently available? Well, aside from the waiting list, Google has some quite generous quotas in place (text from the Google App Engine Blog):

During this preview period, applications are limited to 500MB of storage, 200M megacycles of CPU per day, and 10GB bandwidth per day. We expect most applications will be able to serve around 5 million pageviews per month. In the future, these limited quotas will remain free, and developers will be able to purchase additional resources as needed.

500MB is about half of a free Picasa Web Albums account; for reference, I'm still using less than 170MB of my quota there, even with all the images I've uploaded from Picasa and Blogger. 10GB is a huge amount of bandwidth for one day. Five million pageviews a month is about ten times the popularity of Matt Cutts' blog, and he's fairly well-read.

So how would I use this service? I'm not sure I would right now, seeing as how I don't have any ideas for using it. I've considered Amazon's Elastic Compute Cloud (EC2) service for running my own wiki, rather than just going to a conventional host, but haven't yet figured out how I'd do it. Google's new offering doesn't seem like it would work any better, and might actually be impossible for a while because MediaWiki (my preferred wiki engine) is written in PHP.

Matt Cutts thinks this new service will make it easier than ever to launch a startup; Garett Rogers is of the opinion that Google App Engine is a direct competitor to Amazon Web Services' offerings. I agree with Matt, but disagree with Garett and the other bloggers who think Google's service is comparable to Amazon's services. Google App Engine is designed to be the entire app, data storage, processing, and serving; Amazon offers building blocks you can use individually through your own server or from within EC2.

I find it quite Googley that the basic plan for Google App Engine will be free to use. The current preview limitations will be free even after pricing is announced. Amazon has no free offerings (but thankfully no minimum; minimum charges are anachronisms anyway).

As an aside, jQuery developer John Resig tweeted something funny about App Engine last night:

Haha! You can register apps that you don't own, on App Engine. I now 'own' the official App Gallery and @natekoechley's shoutout app.

That's rather interesting, and might highlight a security problem with Google's newest product. I'm sure it'll be discovered, localized, and fixed in pretty short order.

Also, Jaiku announced this morning that it will be moving to App Engine in the near future. If nothing else, it will relieve the Jaiku team of having to worry about scaling the system. Congratulations, guys!

What do you think of the new service?

[Image composited from two logos found on Google App Engine sites]