rustindy

masquerade: in all my experiments, I have yet to find good CSS-only code to handle vertical sizing that is based on the browser’s size. It’s either ignored, causes scroll-bars even with 0 for the margins and padding, or requires 3 different sets of rules for different browsers. So when I’ve needed to do it, I’ve used Javascript to resize the thing, but leaving the div full-height for graceful degrading.

The limitations of that approach are definitely apparent though. If you wait until everything is displayed before resizing the div, it’ll show up full-sized first (with scroll-bars) before resizing. If you keep the content div hidden until it’s resized, you’ll violate a basic usability tenet (to wit: main content should appear quickly, if not first). And different browsers have different ways of figuring out what the actual dimensions of the available browser height is, so you need to do some client-sniffing (which you’ll have to do anyways if you use the AJAX techniques).

But then, if you’re willing to put in the effort, you *can* do vertical browser-height based sizing solely in CSS, thus avoiding all of the above problems, with the exception, possibly, of having to use a hack or two so that different browsers (specifically, IE and IE/Mac) display the same (the Holly Hack and the Box Model Hack are the two biggies to learn if you’re heading into anything more than beginner-level CSS).

In any case, good luck with it 🙂

The only thing I really didn’t like about it is that it’s hard-coded for something just slightly larger than 1024×768.

The problem with designing a site this way is that you need to size things vertically as well as horizontally. At 800×600 the site is nearly unusable, and because I get a vertical scroll-bar at 1024×768, I end up with a horizontal one as well. Narsty.

My advice? Recode slightly so that Javascript will resize the elements to vertically fit whatever the available window height is. Simple CSS will take care of horizontal resizing.

Aside from the sizing issues, that template is approaching serious coolness, and only a bit of graphic design away from severe coolness 🙂

InnerEyes’ main page is written in ASP actually, not PHP. I know ASP way better, so I can write something in a hour that might take a day with PHP.

In short, and if you don’t mind skipping some of the error-handling with blanket on error resume next statements, the left column of the page is probably about a days work.

The trick is preventing the page from reading each and every blog every time someone loads the page – that’ll kill the server eventually. InnerEyes is set to refresh every 5 minutes – it just stores the last time a database refresh was done in a table, and checks that table every time the page is loaded. If it’s been over 5 minutes, it’ll re-query all the blog databases directly for posts and comments, and it stores all that info in another table. If it hasn’t been 5 minutes yet, it just reads the existing information from that same table.

In theory, a simple thing to do. In practice, you do have to put a little effort into making sure that the information you receive from the blogs is what you’re looking for (so you need to calculate the GMT timezone from the server’s timezone and limit your query so you don’t get future-dated posts, you need to ignore drafts, you need to ignore moderated comments, etc.).

Technically, using the blog’s RSS feed is safer, since it will already have filtered out everything (it only displays published posts, approved comments, etc.), but it’s really slow for more than a few blogs. InnerEyes originally used RSS, but the update for 20 blogs took almost a minute. So I rewrote it to query the blog databases directly and now it checks 75 blogs in 1 second 🙂

There are two ways (that I know of) to do this.

First, use a 3rd party URL rewriter. Sorry, but this is really the best way. There are a couple of free ones if I remember correctly, but to get anything resembling the functionality of Apache’s mod_rewrite you’ll need to cough up a few bucks.

Method number 2 involves a bit of recoding in WordPress. I’ve used this technique several times now with 100% success. IIS can use any file (including ASP or PHP files) as a custom 404 (file not found error, which is what you’ll see if you try to type a “pretty” WordPress URL on an IIS server) page. If a new file can be created to act as the 404 page and pass control to the necessary existing pages as needed, this can work. In ASP, the call is Server.Execute("filename.asp"), but I don’t know the equivalent PHP.

Not a job for a beginner to work on. I highly recommend the first solution.

The third way is simply to get Apache installed on the server. Worked for me 🙂

The Windows user that starts with “IUSR_” needs to be given write and modify permissions to that upload folder. If you don’t have remote access to the server to do it yourself, you’ll need to contact your host to have them do it.

That username is the anonymous web user in Windows. It’s the username that IIS uses to access publicly accesible websites. It’s possible that your host is using a different account for that purpose though – so make sure you tell them to give the anonymous web user write access to that folder, whatever the username is.

Viper, I wasn’t suggesting free hosting on a sub-domain for them, I was just giving them a yardstick to compare against. There are hosts I know of that charge $10/month US for 250MB of disk space and 1GB of bandwidth – that’s insane, and a rip-off.

Sounds like someone changed the permissions on the /tmp folder and MySQL can’t write to it anymore…