/devel

It’s an expected fact of a geek life that interesting technologies and gadgets appear when you don’t have either means or time to play with them. No surprise then that Google Gears was announced exactly at such a time for me. As you’ve probably read elsewhere, Google Gears is a browser extension for Firefox and Internet Explorer (with Safari coming up) which lets developers create web applications that can also run offline on Windows, Mac OS X and Linux.

Google isn’t the first company doing something like this. Adobe has Apollo, Microsoft is working on its own thing, Mozilla added support for offline storage in version 2 of the fox and there are some less well known attempts like Dojo’s. What makes it special are few things.

First, unlike Apollo or Firefox, it’s not a special environment and it’s as cross-browser and cross-platform as it gets these days. Google is also trying to build an industry support for this and Adobe already announced it will support Gears API in Apollo. Same has already been done by Dojo Toolkit. There’s also an intention to make it an open standard by submitting proposal to WHATWG/W3C (and hopefully them accepting it).

Personally, I can’t wait to play with WorkerPool API (which seems to be overlooked in all the excitement). Having a possibility of running time-intensive operations in a background without the fear of triggering “unresponsive script” dialog is a wish come true. Even though you can’t access objects document and window (and hence any part of the DOM). Reason for this is that background scripts don’t share any execution state and hence can’t all access unique objects like aforementioned ones.

This might limit usefulness of the API somewhat, but there are still plenty of uses that come to my mind. What also comes to my mind is a problem, that isn’t really technological. Web applications gave impression to public of being fairly safe to use even on public computers (which isn’t really true) and I fear many won’t understand that new tools may now store private data where they don’t expect or want them to.

It’s getting more and more horrible to travel around with planes. Senseless airport rules or those which benefit only airport employees can still get me angry even days after return home. These days I wouldn’t dare to change planes at Heathrow unless I had at least two hours to do so and would need at least 3 hours not to be nervous about it. Still, there’s a positive side to it. It makes coming through the home door a happier event.

Somehow I don’t mind the actual waiting at the airport. On our last trip we, my wife and I, waited several hours for our next flight which gave us ample time to sit in cafeterias and observe functioning of an airport.

We watched couple of men loading luggage on a plane and unsurprisingly they didn’t handle suitcases gently. Our fairly new samsonite is a witness that this is quite common and I don’t expect it to last for more than a few more years unless we reduce amount of traveling we do. Still, I was astonished to see how far they can throw big heavy suitcases. It certanly can’t be good for their backs.

Looking at flying luggage got me thinking. Wouldn’t it be better if the whole process of loading and unloading bags could be mostly if not completely automated?

It could be done the same way as it was done with ship containers. Just standardize a few sizes in which all suitcases would come and build airplane cargo areas that take those sizes into an account. I believe you could handle most luggage completely automatically and use human gorillas only for pieces that don’t fit.

You could also stimulate adoption with appropriate weight or pricing benefits for those who would use such luggage. This way we’d all win. Or is there something I’m missing?

WWW2007 is wrapping up tomorrow, but today is the last day I am attending. It’s been really refreshing to be at a web oriented gathering without hearing anybody say: “Web 2.0″.

I enjoyed listening to talks and learned that I have to brush up my math and maybe learn a new thing or two. I still learned a lot despite my personal shortcomings.

My original reason for attending was AIRWeb07, a workshop on adversarial information retrieval on web, because I wanted to learn what anti-spam community is working on and what sort of ideas and tools are being discussed.

The answer to an outsider like me seems mainly to be a problem of web spam polluting search engines and algorithms trying to discover spam pages from graphs of connected pages.

It is a very simplified description, probably even too simplified, and I have to say that I admire the work and ideas done and shown at the conference. At the same time I was very surprised that focus has been so much on links and practically none on characteristics of spam pages themselves (with notable exception of a submission by Gordon Cormack). I understand that crawlers are somewhat out of vogue in research community, but it still seemed kind of odd.

It is as if someone trying to know me would completely ignore what I say and solely concentrate on who am I hanging out with and how do I go about it. It seemed to me that the best way would be to combine both, which is what prof. Cormack proposed as well.

In a way I think I may understand this focus. It’s very easy to add spam links on an open participatory web and obviously nobody can control what gets published on all those pages out there. So it’s natural for search companies to concentrate on points that they do control, like their index.

Still, I feel as if not enough has been done on preventing stuff like comment spam. Better tools would only lead to more isolated spam pages and therefore to an easier extraction of spam networks from an index.

Then again, I don’t do this kind of research for a living and may be only talking nonsense.

The other thing that’s been on my mind is a general agreement that this is an arms race and one which won’t be won unless financial incentives driving spammers can be eliminated. Question came up if we should try to break our own protections (as cryptographers do) and how would we fare?

I have no doubt that people I met fighting spam are much smarter than most of their opponents, which is why I don’t doubt that they could do it. I doubt only if they should do it.

It’s not really intuitive that the best thing in an arms race would be to show the other side how to build a nuke. Especially when it seems we are busy enough fighting problem in its current variations. Then again, I’m not convinced it’s a bad idea either.

I won’t go much into other talks, as I feel even less qualified to judge and anyone can get a better insight just by reading accepted papers that are all available online. What I heard or read has been in general of a high quality and if this year is anything to go by, then try to attend next one, which will be next April in Beijing.

By the way, visiting Banff is a must even without a conference as an excuse, since the place and even more its surroundings are just stunning. And if you do decide to go, do yourself a favor and contact Alicyn for a place to stay.