I know it’s only been a month, so it’s obviously far too soon for me to be publishing any kind of a script update. However, these were things which just needed to be done.

First, I’ve greatly improved the internationalism of the script: it now operates searches on UTF-8 encoded data, so it should be great for working with any language which requires characters in that character set. Testing this kind of behavior is difficult, so if you should encounter any problems please do let me know!

Second, I improved the pagination of results. Previously, if there were 150 pages of results, you’d have to see links to every single page of the results. Granted, this can be advantageous, but it’s not really practical. In this updated version, when there are more than 10 pages of results you’ll only have links to the first page, the last page, and two pages before and after the current page. This boundary is configurable — you can offer 50 pages before and after the current, if you wish, but the default is 2.

Last, I changed the way the search query was checked before being sent to MySQL. The original version of this script used my own custom-authored regular expression to secure the request. This was probably very secure, since it functioned by only allowing certain characters rather than checking restricted characters — but it did make things such as multibyte queries pretty much impossible. So it was time to change over to a more standard method. So be it! I got what I needed out of the regular expression; it didn’t need to stay in the script.

Following the publication of a brief article on Search results design by Adaptive Path, I decided that revising my database search script was a valuable goal. Specifically, meeting the checklist in that result was probably not a bad idea!

It’s not that the previous version was terrible, but I knew perfectly well that it could be much better.

The additions to the script are pretty straightforward:

Additions:

• Added: Made row highlighting available in both tabular and list-based search results.
• Added: Search terms are now highlighted in search results.
• Added: The default sort is now to order results by query relevance.
• Added: Paginated navigation of search results is now available.
• Added: Translation base file [English], so translating the script is easier.
• Added: Basic Spellchecking [English]
• Added: Default stylesheet

Changes:

• Changed: Text excerpts are now truncated at word boundaries, rather than in the middle of words.
• Changed: separated results template information into external include files for easier upgrading or modification.
• Changed: Included the search form as part of the script so that search terms could be automatically returned to the search input.

The spell checking is the most exciting addition in my view. It’s hardly complete, but it’s based on a list of 4,068 common misspellings available from Wikipedia. This addition has significantly bulked up the total download size, since I’m including the spell-checking database as part of the download, but I think it adds a lot of value to the script.

I’ve also added a translation base file to the package, to make it a bit easier for users of the script to port it to their own languages. Unfortunately, I haven’t yet had time to seriously work on the internationalization of the search script itself, so (to be entirely frank) this is an area which the script isn’t really well suited at this time.

Internationalization is next on the list, however. It’s a high priority at this point, since internationalization ranks as one of the most reported problems with the script.

With spell-checking in mind, I think it’s appropriate to provide a healthy reminder of the limitations of spellcheck:

Candidate for a Pullet Surprise

by Mark Eckman and Jerrold H. Zar

It didn’t quite take 2 years, at least. But very, very close. The last release of this script was May 25th, 2006 — - so I made it just under the wire. But this is also a bit more than just a script update. In fact, this is a complete overhaul. I’m not certain that there’s actually a single line in the poll script which is the same as the previous version.

This was necessary; because the previous version was, in a word, pretty lousy. It may have acted as a decent jumping off point for some beginning programmers to code their own; but, on the whole, it was not a sophisticated script, and it was nothing like developer-friendly.

This new version, leveraging the power of Google’s Chart API and some clever scripting by Christian Heilmann provides a better end result with fewer potential problems for the user.

I’ve vastly increased the flexibility of the script — - which once could only support a fixed 2 — 5 options in a single question — - to provide support for any number of questions with any number of options, customizable at the question level. I don’t anticipate that anybody will be authoring 100 question polls with this…but they could, in theory.

And, to cap it all off, I’ve added an administrative interface which allows users to perform most of their basic management needs without needing to crawl into the database. Hopefully, it won’t prove to be too buggy.

Are there likely to be bugs in this? Yes! So, if you download this and try it out, please let me know what you notice. I caught quite a few; but I think it’s safe to say that there are a few left in there.

And by “a few,” I mean “actually, there could be a lot of ‘em.”

Check it out or download the package now.

I received an interesting comment from my contact form the other day. I don’t need to respond to it, as the sender left a clearly false email address as their response address, but I do feel that it poses an interesting question for me.

This is the message in its entirety:

Dude how come your example poll have this amazing format, cool graphics, and when I download and installed yours, looks like SHIT?

I mean, thanks for doing this and all that, but come on, you are showing a FALSE example on your web site.

The example poll referred to is this: an example installation of a free MySQL/PHP polling script available on my website. Now, I find it hard to believe that anybody actually thinks of that example as having “this amazing format, cool graphics,” but that’s not really the point: the question is what a “free script” should be expected to include.

This person obviously expected a fully-realized, designed installation. What I provided was nothing but raw HTML and PHP scripting. No “out of the box” styling at all. This is what I generally desire out of a script: if it has a few hooks for CSS and semantic code, that’s great!

I can certainly agree that if you want something you offer to really become a major player in the world of popular downloads, you’ll need to put in a fair amount of work in providing easy template styling, etc. But I hardly think that should be expected for a simple web site add-on.

It leaves me curious: when you download a script, what do you expect of it?

Just a quick announcement that I’ve updated my search script for Boolean matching in MySQL. The latest version is available for download on the PHP/MYSQL Search Engine script page.

The changes to this version include:

• Fixed: various minor bugs
• Fixed: multipage results not available with quoted strings
• Fixed: full-text query problem
• Fixed: some compatibility issues with different PHP/MySQL configurations.
• Added: sortable results
• Added: configurable field label names
• Added: sample results page
• Added: test mode for troubleshooting.

It’s not a massive update; but the script should be easier to use and more reliable.