Showing posts with label google. Show all posts
Showing posts with label google. Show all posts
Thursday, 27 January 2011
Nothing's Perfect — Funny SPAM Detection
I've been itching to share this experience with everyone for a while now, but never got to actually write it. You know, I'm using gmail and I am receiving hundreds of emails (from mailing lists, bugzillas, real people etc.) each day (and for that reason I have a lot of filters to keep them sorted). And a tiny amount of these are SPAM. Sometimes google fails to detect obviously SPAM as such and I can understand that, however it is almost ironic how some perfectly sane and legitimate mails end up in SPAM folder. Ever since I noticed a mail from my thesis' advisor in SPAM folder I check every mail there. And among those that almost 100% end up there while they shouldn't are mails to webkit-dev mailing list from people with @google.com mail address. Funny, isn't it?
Thursday, 19 August 2010
The WebM Problem
Well, lots of people are hyped about the WebM format (which is basically some kind of matroska based [btw. can be (de)muxed properly with new matroska tools] container, VP8 video codec [which was recently opensourced by google and thus is available in fedora] and Vorbis audio codec) which is understandable as it brings open standards to the web and you can play it even with a fresh install of fedora (well, since it was introduced only recently, this is true since the now-soon-to-be-released F14 Alpha).
And to lead the way, youtube is supporting it. But there's a huge but. On youtube it's available only if you enable html5 beta testing, by far not all videos on youtube can be played without flash and out of this subset only even a smaller number of them have been transcoded to WebM, so as a result if you are lucky, you'll play almost anything, and if you're unlucky, you'll either keep getting messages about missing flashplayer or missing h264 codec :( And to add to that, google is being ass and instead of querying the browsers for support, it uses user agent sniffing and thus effectivelly bans all other browsers than opera/firefox/chrome/safari/internet explorer from html5.
Someone should tell the web developers that there are zillion browsers out there that have support for advanced web features comparable to that in opera/firefox/chrome and that user agent spoofing is broken by design (unless you target it the otherway round &ndash send some quirks to browsers that are known to be broken). It's not a hard thing to masquerade my midori as firefox or safari, but I would prefer if I hadn't have to do that…
And to lead the way, youtube is supporting it. But there's a huge but. On youtube it's available only if you enable html5 beta testing, by far not all videos on youtube can be played without flash and out of this subset only even a smaller number of them have been transcoded to WebM, so as a result if you are lucky, you'll play almost anything, and if you're unlucky, you'll either keep getting messages about missing flashplayer or missing h264 codec :( And to add to that, google is being ass and instead of querying the browsers for support, it uses user agent sniffing and thus effectivelly bans all other browsers than opera/firefox/chrome/safari/internet explorer from html5.
Someone should tell the web developers that there are zillion browsers out there that have support for advanced web features comparable to that in opera/firefox/chrome and that user agent spoofing is broken by design (unless you target it the otherway round &ndash send some quirks to browsers that are known to be broken). It's not a hard thing to masquerade my midori as firefox or safari, but I would prefer if I hadn't have to do that…
Tuesday, 9 September 2008
Some WebKit news
Ok, with the recent talks about brand new browser by Google, I think I can post some "status-update" on the state of WebKit-gtk. Why? Because google chrome uses WebKit as its rendering engine - and that's very good news. With Google being directly interested in WebKit development we could say that bright future awaits. Google uses it, QT4.4 has it in, now we only need to implement it in gnome.
Among various speed and stability improvements there is one big news for me (at least) in the WebKit bugzilla. Patch [1] for correct font fallback (so far only for freetype backend) for the gtk port of WebKit [2]. I've built a RPM package [3] with this patch included and the result is great - I can finally read Japanese with WebKit.
One of the worse news is that it seems Epiphany 2.24 will still use gecko, well, what can I say... Take your time and do it right and I'll be happy to wait (and test and report bugs) :-p
[1] https://bugs.webkit.org/attachment.cgi?id=23155&action=view
[2] https://bugs.webkit.org/show_bug.cgi?id=16792
[3] http://mso.fedorapeople.org/packages/RPMS/i386/WebKit-gtk-1.0.2-0.15.svn35913.fc9.i386.rpm
Among various speed and stability improvements there is one big news for me (at least) in the WebKit bugzilla. Patch [1] for correct font fallback (so far only for freetype backend) for the gtk port of WebKit [2]. I've built a RPM package [3] with this patch included and the result is great - I can finally read Japanese with WebKit.
One of the worse news is that it seems Epiphany 2.24 will still use gecko, well, what can I say... Take your time and do it right and I'll be happy to wait (and test and report bugs) :-p
[1] https://bugs.webkit.org/attachment.cgi?id=23155&action=view
[2] https://bugs.webkit.org/show_bug.cgi?id=16792
[3] http://mso.fedorapeople.org/packages/RPMS/i386/WebKit-gtk-1.0.2-0.15.svn35913.fc9.i386.rpm
Subscribe to:
Posts (Atom)
