Video

2.500 lines of code (really smart code!) later i'm getting closer to making this real cool thing, so wallrush is still on a standby. However a friend had a cool idea on how to change the links to the videos into thumbnails, and I already got that - what was it called? - really nice converter somewhere on the server. I suppose 2 hours of playing should fix that.

I'm starting a new job this week, and I'm beginning at economics at the University of Copenhagen next week - last week I moved from Østerbro til Islands Brygge.

Byte 16314

I have made some additional changes to WallRush.com, among other I'm slowly working on showing who added the material. No rush though. Also I changed the way hits on vidoes are weighted against the freshness of the video, so that, hits counts 30% and freshness 70%. This seems to be a reasonable ratio, but it might turn out that 80-20 is even better.

I also remade Qys.dk entirely moving the Quick ASP Upload site to its own subdomain. I also had a glance at what people have uploaded the last 2 months. One word: Insane.

As a last note; is it possible to have a nightmare about an access denied error when trying to write to a file at byte 16314? The answer seems to be yes.

Parsing xml

So I have updated the video add section so that videos are automatically resized to 450 * relativeheight. This required parsing of the html/xml the user supplies, and I chose only to allow, attributes and tags that are needed to show the embed (embed, object and param tags).

I am going to mess around with this a bit in order to make sure that hostile input is filtered correctly.

Annoyance

Yes I am very slowly developing WallRush, but I guess there's some progress. I have finally cracked the RSS import problem - mostly. First of all I wanted it to be able to handle both normal RSS and ATOM. This isn't a crazy thing to demand, so I naivly thought i could use the XMLDoc object to parse both formats. I was wrong - some feeds are invalid and there's no way to make XMLDoc read invalid XML.

Workaround?
It seems, that microsofts implementation of the xmlhttp object believes that it should always use UTF-8 encoding if the mime-type is text/xml, when the responseText is accessed. This is bad, because ASP doesn't - to my knowledge - let me encode text as I please, and lets pretend i might losen up a bit on the parsing of invalid XML by writing my own parser. In that case I have to work with malformed text when using the responseText property. I tried to use the Text property of the responseXML, however it failed in the same manner as responseText returning malformed text because the it was incorrectly encoded with UTF-8.

Why can't i just use responseXML and the XMLDoc object? Because it refuses to parse invalid XML, and I can't expect everyone to pass valid XML every time. I'm of course not the only guy with this problem and it seems that using ADO to fetch the feed, it might be possible to decide the encoding of the text, but I will have to look into that.

However not all is bad. The import feature works again, just not with foreign character sets (like danish). I fixed the object_hit function so preferences is now recorded again for videos. I also changed some of the graphics. I think it is much better now. And i added a list of available videos so that i don't have to go through them all to find those I am interested in.

So bad day today, I'm having the absolutely worst hangovers since that incident back at that bar in yeah whatever - massive pain.

Okay so i spent most of the day trying not to feel my head too much. But now i guess its time to do some work.

I saw another site slightly like mine called wallpipe.com. It looks nice, but it still suffers from the twitter / facebook problem of me being spammed with irrelevant comments from people i only halfway know.

I am quite worringly now starting to use wallrush.com though there's hardly any content. Specifically i feel a need to be able to browse through the videos so this is what I am implementing today. Maybe I'll also fix the remaining add content sections and change some of the design a bit.

Favicon.ico

Yeah well I'm making some sideprojects for real green cash, so wallrush.com is not developing as fast as I desire. However last night I finally made some progress, so now there's a script for importing the title and favicon of linked sites. I think that's kind of cool. Next up is improving the RSS importer, finalizing the sign up, and add pages, and making a better design of the wall.

/Mathias

PCA

Some time ago I read a very interesting article, about a competition by netflix.com (read more) where the point is to make a program that can predict users opinions of a new movie based on previous judgements of other movies.

The authors of the article had noted that no matter which method was used for prediction none of the competitors could get closer than 90%-95% (i think) accuracy in their estimates. The really interesting thing was that though very sophisticated mathematical methods yielded better results it was about 2%-3% better than simple methods.

Obviously this means you can use a lot of time being slightly better, but you should consider if it is worth it. That might be the case if you want storm predictions or something like that. But for a "simple" social entertainment site 90% is properly going to be as good for at start as 95%.

So I am opting for a "simple" method of deciding wheter content is relevant for the user. More on this later - gotta feed my self :)