I’ve been hanging out on #kde-www for a while, and today we had a real meeting “to determine some future goals for the porting of kde.org and its sub domains over to a CMS system and also on the future of open collaboration services within the KDE online community”.

Read more ...

Trying to get my custom 404 page to work (yes, it works now) once again, I ran once again into the thickets of the Apache server configuration. Try as I might I didn’t know how to get it to understand that, yes, I have a .htaccess file and yes, I do want you to read it and do what it says. I found myself running in the exact same circles as the first time around.

Read more ...

I’ve been wanting search functionality for the website as well as the blog for a while, and I couldn’t get the blog search widget to work on other pages than blog pages— after all, I don’t know any Perl yet. I was just about to teach myself Perl in order to write a search widget when I happened to mention it in alt.html and someone suggested using Google Custom Search or Atomz.

Read more ...

Now, of course, I find that the design breaks anyway in Firefox because it doesn’t like non-floated and floated pictures side by side. And I don’t want to make the text and the picture one picture, because the text should stay in the middle of its box and the picture shouldn’t leave its position on the right.

And why the text heading shows up with a grey background is a mystery. [ETA: that mystery solved: somehow blosxom thought the heading of the previous article was part of the header; when I let the header inherit the body background, as it ought to, the text heading inherited the div background. Ah well.]

Back to the drawing board once again.

[ETA: made the text-like object float left as a kludge. Don’t make your window too narrow.]

Several people told me that the letters of the subheading were falling out of the header, especially with larger text and/or narrow windows. Making it specialised <h1> without bottom padding and <h2> without top padding instead of <h1> with <br> and a smaller-type <span> fixed part of the problem— Konqueror and Firefox keep <br>-ed text together, with the padding all around the whole thing, Safari puts padding between the lines too. But still the tail of the ‘g’, and especially of the ‘p’, crept into the article space. The header needed to be 200px high because of the picture, so I couldn’t very well let the padding take care of it.

“Make the text a picture too,” my other half suggested. I must admit that I’d thought of that, but I’m usually against pictures-of-text where actual text will do: I feel it’s something that bad (incompetent, not evil) or lazy web designers do. But after some thought I gave in: the heading is now a picture of the previous heading at its best, set on a background of the same colour (#d2cfd8, which I called “ingridgrijs” in kcoloredit), centered in its <div id=”header”> so it stays neatly in the middle on a wider screen. The overall design still breaks at large font sizes —the text runs into the sidebar— but I’ll fix that when I run out of other things to do.

(Also, if you don’t have URW Chancery or Zapf Chancery but rely on your generic cursive font for text headings, the font won’t match; sorry about that. I’ll fix that when font-embedding technology comes of age.

It all started with the picture at the top right. One of the students in the art class I was sitting for made it, and as I was taking a photograph of it I thought “I want that as an avatar!” But, come to think of it, I don’t currently do much that I need an avatar for, so I’m using it for the blog. It looks much more like me than any photograph I can find: my back and shoulders are quite unmistakable. Thanks, Ingrid!

This morning, instead of going to church —because it was clear that the family cold had caught up with me, my throat didn’t want to sing and my head didn’t want the company of lots of people— I sat down and revamped the whole thing, as I’ve been wanting to do for a while.

I’m still working on a calendar (I do have a plugin that sort of works, but it refused to go back neatly to the page it came from so I got URLs like http://www.valdyas.org/foundobjects/index.cgi/2007/church/2006/, and generated single-article pages without the breadcrumb trail that I’d just been tweaking to look the way I wanted) and on an atom feed to keep the outdated rss 1.0 feed company. Also, I haven’t done the comments page yet, but <sob> I get hardly any comments anyway.

(And just now my other half shows me the page in Safari, where the sub-heading falls out of the header; back to the drawing board!) [ETA: fixed that —the fix is more elegant anyway]

Note: the blog is optimised for Konqueror only. It’s such a personal thing that I think I’m entitled. If you’re an IE user and you think it looks rotten, I recommend upgrading to Firefox; that doesn’t have as pretty a cursive font as Konqueror or Safari, but it was good enough when I looked at it.

For reference: Petrus en Paulus nieuws. If you want to follow what I’ve been doing you may want to open it in a new window, but that should be your choice, not mine.

A while ago the webmaster of another parish sent me, as webmistress of our parish, his changed link. I took the opportunity to check everything on the links page while I was working on it anyway. Several parishes, for some reason mostly in Belgium, turned out to have their website set up as a blog. At first it seemed strange, but it sort of grew on me, and eventually I thought “I can do that too!”

Read more ...

Someone in Romania googled for ‘lionel rychie’ and, because I have a Useless Blob called Lionel (warning: he’s jumping up and down), and there’s someone called Rychie Korbes in Hans Brinker or the Silver Skates, and valdyas.org has a whoppingly high page-rank, the third entry on the results page pointed to the Hans Brinker names deconstruction.

Which the person from Romania proceeded to let Google translate into Romanian.

I happen to be able to read a little Romanian— just enough to see that the English-Romanian Google translation is as hilarious as the Dutch-English one, but not enough to actually correct anything (as I’ve done with other pages of mine that I noticed people had had translated). The universal translator is a long way off yet.

Google does ask “Did you mean: lionel richie” which is probably what the person from Romania was looking for.

To get Rychie Veray (a town: Veray-on-the-Rycha) instead of Rychie Korbes (a girl probably called Rietje) one has to search for ‘lionel rychie veray’, and there’s only one hit for that— well, four if you let it show all the variants. Which is strange, because the Rychie Veray post is much more recent and has been read much more.

… what if all those people who have been finding my web pages by searching for linguistic terms (“nouns and adjectives”, “noun classes”, “relative and reflexive pronouns”) really think it’s an official, scholarly linguistics site? About a language spoken by real live Earth humans? Should I put a disclaimer on every Ilaini page, making it clear that it’s an invented language?

It’s probably an effect of our absurdly high page-rank, which is itself an effect of my other half’s exploits in the KDE community. Whatever people search for that happens to be somewhere on our pages, valdyas.org is likely to turn up on at most the second page.

With the new web design it was easy to stick a statcounter on every page. I’m surprised that my website gets about three times as many hits as my blog (but that won’t keep me from blogging).

Read more ...

It’s no longer an old website with some new material, but a new website with some old material.

Today I did all the roleplaying writeups and some of the Ilaini stuff, strewed the site liberally with redirection pointers (not nearly enough, I fear; what I want is a couple of custom 404 pages but I can’t figure out how to make the apache server believe that) and made a sitemap with everything already updated or, like the Purplish Cooking Pages, not to be updated any time soon. Then I checked all the links and permissions in my local mirror and —wait for it— removed the old page before putting up the new to avoid duplication.

38 folders and 290 files instead of the previous 41 folders and 345 files. I did notice the discrepancy, but thought that some of the files had still been in their old place as well as in their new place.

Then I went to fix all the permissions that fish had apparently screwed up. And found that all my roleplaying writeups were gone, which I’d just been spending half a day working on, and worse, of which there was now not a single copy within my reach except on paper. Somewhere along the line I’d copied an incomplete version over a complete one.

But Google was my friend! I could salvage everything from its cache <whew>. Now I only have to do the restyling again and write a new index page, which is a doddle compared to what it could have been.

I’ve started refurbishing my web pages at long last. Already done at the time of this post:

• home page
• FAQ (also with some new questions and answers)
• Valdyas index page
• the fun grammar-example pages:
• The old woman without a stick
• The people who didn’t go to the market
• The prince who turned into a frog

I’m learning a lot of new CSS tricks. And, if I may say so myself, it’s pretty. It does make me think that the blog is too blue, but I don’t want to think of a new style for it while I’m constantly making little tweaks to the style of the web pages.

These days, now that everybody in the house can understand English, we don’t watch English films with Dutch subtitles so often any more, but whenever we do there are always a couple that make me shake my head and say “I can do better!”

And I got a chance to do better: two videos made, and subtitled in English, by a friend who works at a university library in New Zealand.

It was fun, easier than I’d thought (dotsub has a good user interface, though it’s confusing that you get to the next input box with Enter, not with Tab) and I think they’re pretty decent. I want more! If anybody reading this has a video in English or German that needs a Dutch translation (or a video in Dutch that needs an English translation, for that matter) don’t hesitate to drop me a line.

Time for another roundup of search terms.

(apropos of nothing, I seem to have about 30 regular readers)

Read more ...

…only not just now. But I was carrying the old laptop I was using (which used to be Secunda’s, and my other half’s before that, and his employer’s before that, an ancient but still serviceable little Gateway) up the stairs because I’d just copied some files on to it —no network reception on the roof terrace with the almost equally ancient wireless card— when I slipped and had to let go of one corner to keep myself from pitching down, and that corner banged on the top step.

Read more ...

Last night, while we were immersed in roleplaying, our ADSL connection stopped connecting. The girls noticed, but thought it was a dip of the kind we’ve been having all too frequently and it would be up again within minutes, but they were going to bed anyway so they didn’t try, or think of telling us. We didn’t notice until the other player had gone home and we wanted to wind down, check mail, talk to people on the other side of the world, and in my case make a start on writing the writeup. My other half tried to fix it, but it wasn’t one of the simple things (rcnetwork restart, try to ping, OK, works) so he said he’d try again in the morning and for all he knew it would be up again then.

Read more ...

I’ve now had a counter on both Found Objects and the church pages for the whole month of April. I know our server must have stat-counting functionality somewhere too, but someone pointed me to Statcounter and that works, it can be invisible, it’s free and easy to use, so I’d rather be lazy and let something else gather the information for me.

Read more ...

Comments are working again; it was a permissions problem after all, but not where I’d looked for it. Boudewijn fixed it (thanks!). Of course, then I had to go and break something else, but I could fix that by myself.

We now return to our regularly scheduled program. No more tweaking or rearranging for a while, especially as Holy Week is about to descend on us, and we have house-guests now and will have more from Wednesday.

I’ve overhauled my categories. All the old posts are still there apart from two outdated announcements. With the right date, too, because I discovered cp -p. It may cause flooding of your RSS feed (though I hope preserving the dates will prevent that), for which I’m heartily sorry, but my blogging habits have changed and the old categories didn’t work for me any more.

Internal links will work again when I’ve overhauled those too; that may take a while. [ETA: now done, it was less work than I’d thought because I could at least semi-automate the searching]

If you’ve lost the link to a post and can’t find it even by searching, please pipe up and I’ll tell you where it is.

Don’t send from hotmail if you can help it. I don’t know why my spam filter thought the header was forged (for all I know it was perfectly genuine) but I suspect (some of?) the hotmail gateways assign random IP addresses, without having an actual host with that address.

Also have an address without numbers if at all possible. The “likely spammer email” mentioned below was of the type “name2006”. This only got 0.4 points, but it was the straw that broke the camel’s back.

If you can’t help sending from hotmail (for instance, as in this case, because your message is the change-of-address from hotmail to something else…) try to make it send plain text. My antispam software is very suspicious of HTML. I’ve already deleted my own filters that throw away all HTML to get rid of some of the false positives, but SpamAssassin is diligent.

I don’t know the reason for the 40% to 60% Bayesian spam probability. Perhaps because there was very little text, only “Hi, my old address is X, my new address is Y, thanks”. But that didn’t give any points, so never mind.

I did catch this one, but I can’t guarantee that I always will. I have SpamAssassin and the filters to avoid having to see every single message that comes in.