Torrential Web Dev

I've just recently come across an interesting tool available at seomoz. The page strength tool is designed to give a more thorough analysis of the importance of a page than the google pagerank.

It looks at a host of factors, including the google pagerank, to give a more detailed and more up-to-date analysis of a page. It takes a bit longer to gather the data (there are more sources to query) but I think it is a more useful tool than just looking at the google pagerank.

There is an interesting article at snook.ca looking at adding an image to an rss feed. RSS feed aggregators tend to strip away the branding you would find on a website to the point where it can be hard knowing where posts are coming from.

The post looks at two ways to maintain some of the branding, a favicon, and then in more detail, an image. Adding images to both Atom and RSS feeds is covered in detail.

Overall a must read for anyone whose audience makes significant use of RSS feeds.

This entry is one of a series looking at fetching contacts. For a summary of all the important facts, a demo and the download link please view the MSN contact grab page in the scripts section.

Unfortunately I have received reports that this script is no longer working. As there is now a better alternative available directly from Microsoft I currently have no plans to update this script.

In my rush to get the announcement out about the web service finally being up I completely forgot to mention an important point. Namely that if you do use the service then the MSN login details for any users of the service will pass through my server. If the correct log in details are supplied then I'll also be returning the email addresses for all their contacts.

As such if you have any concerns regrading my reliability you shouldn't use this service. I'm not going to keep any email addresses and I'm certainly not going to keep any MSN login details. Having said that though it is a question of whether you trust me to keep my word. Hopefully you do.

Well it has been quite a while since I posted about my intention to create a web service for my MSN contact grabbing script. Unfortunately other commitments have meant I haven't been able to spend any time on this project until today. I've quickly thrown the backend and a class to interact with it together and now here it is.

Continue reading "MSN contact grab web service is finally up"

Via Ben Ramsey's blog I've come across an interesting post discussing what is possibly a better way to teach PHP. Rather than giving the reader the bare essentials to accomplish the task they wish to undertake it is suggested that the focus should be more on what is best rather than what is fastest/simplest. Many books focus on putting together simple scripts and skip over the finer points of validating user input and other security matters. It is an easy trap to fall into and although it may produce quick results does it produce more knowledgeable php users? I have to answer that in all likelihood it doesn't.

Ben Ramsey raises the interesting question of whether a book focusing on producing better php programmers, rather than putting together a site quickly, would actually sell. In his own words,

While I agreed with the concept in general, I questioned the marketability of this approach. When a person picks up a book on PHP, I asked, what is it they want to do? The simple answer is: they want to create a dynamic Web site, and they want to do it now.

Although I doubt this is true for all aspiring php users it probably includes the majority. On the other hand this is a vast market so maybe there is room for an alternative approach. On a similarly positive note it could well be a book that more experienced programmers could recommend to anyone wishing to learn.

I seem to be getting a lot of visitors to my post presenting a script for grabbing an MSN messenger contact list. There have also been plenty of comments which is again good news. They have however brought up a few issues. Continue reading "Development on the MSNM contacts grabber"

---

A little while ago I wrote a tutorial looking at how to grab the contact list of a user and then send out an email to each of their contacts. The scripts in the tutorial did exactly what they claimed to do but they were fairly limited. Since writing the tutorial I have come up with a few ideas on how to make it better. Continue reading "Follow on to contacting a contact list"

I've just noticed this post discussing one of the classes in the phpclasses August awards. The image nudity filter class came in second dispite being 'extraordinary'. I remember giving serious thought to voting for this class but I chose not to based on some negative reviews. Those reviews have since disappeared but with the increased discussion I felt the class deserved to be re-examined.

There is an online example so the class is easy to try out and as such I spent the last few minutes playing with it. Despite being voted second I have to go with the negative reviews, the class seems to simply identify skin tones. This means that a solid block of colour, essentially a blank image, returns a result as 100% and an image containing nudity will return a very low score as long as the tint is slightly adjusted.

This class is certainly 'innovating' and a significant step in the right direction but it is not to my mind 'extraordinary'.

After writing the article looking at fetching the google pagerank I realised that I never did add any detail to my assertion that caching the value is quicker than querying the google site every time you want to find out what the pagerank for a site is.

To test the speed of using the cache over connecting to google I looked up the pagerank for 10 sites 4 times each. Each of the 4 queries were done without clearing the cache so the google servers were queried once for each site and then the cache was queried 3 times for each site. The average over the 10 sites for fetching the pagerank value directly from google was 0.075146 seconds. The average over the 30 queries of the cache (3 queries per site and 10 sites) was 0.002217 seconds.

That works out to be just over 33 times faster. Not at all bad.