Hi there! This is my old blog, I don't hang around here much.
You can now find me trying to mix things up here:
Visit WoodyHayday.com Subscribe To My Email List
Its big and its not clever, but it will tell you how much your spending per year on domains based on a column of them:
=IF(ISERROR(FIND(“www.”,BB17)),0,(IF(ISERROR(FIND(“.co.uk”,BB17)),0,2.99) +
IF(ISERROR(FIND(“.com”,BB17)),0,8.99)+ IF(ISERROR(FIND(“.org”,BB17)),0,8.99)+
IF(ISERROR(FIND(“.info”,BB17)),0,2.99)+ IF(ISERROR(FIND(“.net”,BB17)),0,8.99)+
IF(ISERROR(FIND(“.me”,BB17)),0,15.99)+ IF(ISERROR(FIND(“.eu”,BB17)),0,8.99)))
Working with integration of Facebook Connect for authentication is mostly easy at the moment, once you have got your head around FBML etc. However as they are in between switching from the old Restful API and the new graph API there remains the odd little thing that has only a splatter of useful documentation online. The first one of these I have come across is trying to find out if a user (who is logged in to your site via facebook connect) is a fan of page X.
There is infinite uses for this – with facebook more and more integrating “likes” with their search and other stuff you will need this functionality to drive traffic from facebook. Anyway this used to be the easy case of using https://api.facebook.com/method/pages.isFan?access_token=… to get a 1 or 0 depending on whether or not the user had “fanned” the page/website/whatever.
But then came along the big bad Graph API, which to get this IsFan functionality is a bitch. Getting the “likes” directly works for the page owner (which confuses everything) but not for third parties, showing just the popular/main category “likes” against them (and you need extended permissions.)
Something led me to happen across the following video from google I/O 2010 which gives a full expo (57 minutes) on their BigQuery and Prediction API’s. You can see they have made a lot of ground with ridiculous processing, claiming to be able to run SQL queries (or a derivative) on trillion’s of rows in seconds, essentially for free. BigQuery lets us, the general techy public do that, cool eh? Well yeah it is, but I cant help but ask the cynical, why? Data, I would presume is their goal, and if it’s yours to share and insensitive, BigQuery is probably a BigWin. Otherwise its often we find ourselves as technological problem solvers seeking services such as this, a way to fix X, a quicker way of processing Y, a ladder Z to climb fence A to get to B, which can lead us blind with determination down paths, data under arm, paths that are not as safe, sensitive or secure as we thought.
A short while ago I read a press release about a software capable of producing 3d models of moving characters from a piece of video, I was certain the link was on reddit but I cannot find it. Anyhow it got me thinking about the uses for this, should it progress to a viable standard and there has got to be a few notable ones. For the film industry it will no doubt aid the allocation of 3d work over digital film, although it will have to get slick to replace the use of bright colour identifiers in place of actors.
*update: I found a similar software here on new scientist, its not the same that inspired this post but its relevant, video from there.
So I happened across this post on foreign exchange arbitrage and got distracted for 4 hours into working out if there were any opportunities for using arbitrage betting from here in the uk, what with online gambling being so big and tennis being a simple outcome game I thought I would start with the tennis and see where it took me. More an excuse to practice my rapid application dev than anything else.
If you skip to the end of this post you can see the outcome, or read on, I will be brief.
So I started in excel, found a nice summarising scraper site which gives you all the odds from different bookies in one place (how convenient – I bet they make good affiliate revenue) and did the maths. Within 4 games that evening I found one which worked. If I had placed bet’s across 2 bookmakers totalling £1000 at that moment I would have made a profit of £10.74 (1.74%), guaranteed (unless some natural disaster killed both tennis players?!?)
As I said last post if I couldn’t find a reddit wallpaper downloading program I would make one. Well I found upboat (which is good) but with windows 7 (or display fusion) offering cycling desktop backgrounds every x minutes I thought I might write a more brute force downloader. So I did.
Reddit Wallpaper Grabber is essentially:
…a small program (77kb) that will grab wallpapers straight from their listings on reddit. With it you can download (in bulk) hundreds of the most upvoted wallpapers with the click of a button to a folder on your computer.
It’s only about 8 hours work, is unofficial and probably mildly buggy, but it works for me (and probably everyone on windows > 98.)
Features
If you want to suggest feature requests/point out bugs then please do on the comments of this post. If you haven’t already – go ahead and download it.
Download v0.1 now of Reddit Wallpaper Grabber (its free!) |
|
|
|
Hey blog, it’s been a while eh? Hows things?
May was a month of brick lane curries, a new art exhibition (silent city), mini golf and a bunch of other blurry stuff.
June was a good month (on the whole apart from hackers and google caffeine?!?), did incalculable amounts of coding, new site rollouts, tennis, monopoly, booked new york flights and some other stuff. Wrote a lot, drank a lot of wine, updated a bunch of auction2post sites because eBay updated their api (for the better) – standard summer month.
Back to it I thought I would write a quick post about something which seems to be helping my ebay partner network revenues – that is campaign splitting. I am totally unsure of how ebay partner network campaign splitting will work for you – I am only just starting to test it my self – but from the first week it appears to be helping my largest site overall.
If you are an EPN (Ebay Partner Network) affiliate you will be familiar with their latest scheme which has been in play since Nov 09 or similar – the idea of a quality based click through deal whereby the higher the quality of traffic you send through to them the higher they pay you per click. The hybrid system encapsulating the benefits of PPC while associating to a finer level with quality seems to be working well on the whole but does change the game slightly.
The set-up before the move to EPC was clear – a campaign and its sub id’s (channels, custom id’s) were there to allow the publisher a clearer view on what was working – a means to allow them to refine their sites and provide a better quality through to ebay which returned more sales and which ultimately benefited both parties. This would mean traditionally a campaign per website and sub-id’s (channels) for each sub category, sub page, sub item or whatever would sufficiently track transactions and allow for later refinement, later. However with the adoption of EPC the campaign gained a new association in the relationship.