Web App: Juvely

Well I had a good idea for a post today about the problems of working with perfectionists and how to solve them. I have recently been working on a group project and one of the people in our group is a perfectionist, so I thought I would share my thoughts on this. I needed a few ideas, so I decided to search the world of blogs and see what other people have blogged about on the subject. Unfortunately that didn’t turn out to be as easy as I thought it would be…

A few years ago I wrote a search engine for a friend who had started his own blog - this was at the time when only the elite had blogs and before Word Press was the thing to use. The search engine I made wasn’t brilliant, but it did the job fairly well. It ranked articles based upon how many occurances of the search terms there were divided by the number of words, and how many times it occured in the title.

Last week I started with the search engine for Juvely, I based it upon this, because this is what I had known and loved. After a while I decided to start from scratch because it was too limited, my useful old search engine didn’t even handle grouping terms by quotations. :p The new version uses basically the same SQL queries, but the idea of ranking that I used before also didn’t work for this situation, instead I opted for a ranking system based upon the number of results that link together (it sounds complicated but it makes a lot of sense if you understand how Juvely works).

Now, back to my problem. My blog search for information about perfectionists turmned up appaling search results. I tried searching on a certain web search engine for more information which seemed a lot better, but it wasn’t really what I was after - I wanted to see what other people had blogged about the subject.

I was rather appalled at how bad the results I got from the blog search engines were so I decided to try a little experiment. I guess the reason why I didn’t get much information about perfectionists was because it isn’t a very blogged about topic, so I choose a different one and decided to see how a number of web and blog search engines compare.

A highly blogged about topic was the Google takeover of YouTube, so I decided to use google bid for youtube as my search query. If we do a search on three major web search engines (Google, Yahoo, MSN) we get good matching results. Now, if we try the same with blog search engines (Google Blogsearch, Technorati, Bloglines, Sphere) we don’t really get any good matches. On Google Blogsearch the first result is about a presidential bid on YouTube, on Bloglines why YouTube is struggling, on Sphere a load of links, and let’s not even get into Technorati.

The problem seems to be that most of these search engines rank newer posts much more important than older ones, even if they are completely irrelevant to the topic being searched for. Maybe this works for most users, but for me it didn’t. So i’ll go back to the title of this and say that there is still room in this cramped Web 2.0 for a good blog search engine, and yep, if you are really good you can even beat Google at their game.

Although I may understand why I didn’t get any good results, it still doesn’t help in my quest for information. So if anyone knows of a good blog search engine (or just what I am doing wrong) post a comment!

Hello again everybody!

Appologies for not posting in the last two weeks, but I have been busy baking and James has been busy, well I don’t know as his internet connection at uni is still playing up.

Anyway, I was looking for ways to improve my MVC system so that it could handle complex SQL queries and I came across another MVC system called Cake PHP.

I thought I would give it a go quickly, and wow, I was impressed. It is near enough the same as Ruby on Rails, and in some ways better (althought I am sure the diehard RoR fans would disagree). Deployment is easy, and unlike RoR doesn’t need you to fiddle with anything complicated on your server. Plus it is PHP based which almost all webhosts support, so should run near enough anywhere. The main advantages over my MVC system are that it includes a proper Active Record, caching, probably better security, and tonnes of other things. I decided rather than trying to reinvent the wheel, I would ditch my system for one which lets me get on with programming Juvely, which is now a piece of Cake.

Other things that have been useful are the great documentation. It doesn’t go too indepth in topics (which is good for the begginner) but explains more than enough to get on with your first app. There is also the so called Bakery, a wiki that contains “everything CakePHP”.

The engine itself in my opinion is much more user-friendly that Ruby On Rails, the error messages are beautiful. They don’t contain anything unneeded and clearly explain what the problem is, and even suggest a solution. For example, say you go to an URL for a controller that doesn’t exist.

Ruby just spits out an error message saying the route was not recognised.

Cake on the other hand spits out the error message, details of how to customise the error message, and what you need to put in what file to fix the problem.

As with everything it also has its disadvantages. I spent ages trying to figure out why my HABTM relationship wasn’t working, eventually to find out I had my array in the wrong order. Also the main site for pluggins and addons is by default in French. In order to view the English version you need to register (in French) and then change your language preferences, it would have been much nicer just to have some language buttons somewhere, but heh.

Anyway, CakePHP looks very promissing, and I expect to see a lot more of it in the future! After being public for under two years it looks like a lot will come out of it!

Only one lecture today so I took a bit of time to put together myself a new clock.

The clock images and javascript is based upon this, the rounded corners are created by Spiffy Corners and the rest is done by me!

Get it here!

Enjoy!