WordOwl

Apologies for the cruddy title but I really couldn’t think of a better one, but I did only get up in the last half-hour.

Today’s comics are updated, plus eagle-eyed readers will notice a new comic: Sporkman. Night before last, John Troutman - artist behind Flat Feet & High Heels, as well as Sporkman - dropped me an email, asking if I’d like to take on Sporkman here at WordOwl.

I did ask a couple of questions, and John fired back some really useful answers to them, the most important of which is that I have a colour reference for just about every character in the strip, so I can colour up the strips nicely.

So as of today Sporkman has joined WordOwl.

I’m annoyed at myself I only got 17 strips in yesterday, but there is a reason behind that. As I go along transcribing, I periodically update the development server with the latest strips, make sure everything works how it’s meant to. And for some reason the entire character searching system was broken on Sporkman.

To give you an example, I have 17 strips, 14 of which feature Sporkman in some capacity, and yet searches for Sporkman himself were throwing out all 17 strips.

After routing through thousands of database records, hundreds of lines of code and writing new debug code into it all, I finally figured out what was going on.  For those of you who understand SQL, the problem was a JOIN between two tables that wasn’t specific enough.

In English, I have a list of characters and a list of text entries that they say. The name entry is the same as what you see on the search page (a name, possibly just the first name in some cases). But if you have two comics with characters with the same name, there is going to be cross-over: I have characters called Amanda and Rob in both Sporkman and Flat Feet, and both were getting messed up because of it.

Once I’d waded through all this for about 3 hours I finally fixed it, but it meant that was 3 hours I didn’t get to transcribing. I’ve also been looking at upgrading Sphinx now that the final release of 0.9.8 is out, just to take up more time.

So all in all I’m a bit annoyed at myself for not getting as much done as I would have liked.

Lastly, I’d just like to throw a shout-out to Blank It, one of the youngest strips in WordOwl. Created by Aric McKeown and Lemmo Pew (of Ashfield and Lethal Doses fame respectively), Blank It is a nicely refreshing comic. Lemmo, the artist, said some very nice things about WordOwl in the blog, which I really appreciate. It’s stuff like this that makes me feel good about working on WordOwl - it is coming from the creator after all. So, everyone who has a few minutes for some good comic humour, please head over to http://blankitcomics.com/

If I keep getting praise like that, I might even end up making a Praise page.

Plan of Action today is unfortunately not comic related. A small dose of shopping followed by a larger dose of visiting family.

Just a quick update. I’m still waiting for confirmation regarding permission for the next comic, but in the interim it’s made me realise that there is something I needed to do.

Last night I implemented one of the last features I intended to implement in code for Word Owl - support for characters present in a strip without saying anything. The data is being recorded for all new strips, and we will be going back to collect it for all other strips as part of a proof-reading exercise shortly.

What I have realised is that I didn’t have any support for non-English languages at all. Until now it’s never been a problem; I’ve just de-accented characters before, although the instances are still in the low double digits in total, but the next strip I’d love to add does have French in it. And that means accented characters.

So finally, I’d added in some support for accented characters. Strictly speaking it’s not true support, but in reality it makes more sense than ‘true’ support. Internally, accented characters are re-routed to non-accented ones. Nothing special, of course, but it does equalise accented words and makes them searchable without any further work.

The first actual use of the new software has been with Bruno, actually. The use of the word résumé - with the acute-accented e, in particular. It’s routed through to its unaccented equivalent, resume.

Although this will through out duplicates (instances of the verb, rather than the American-esque noun), this seems better than insisting on an accented version in order to actually find it.

In terms of development, other than new strips and new comics, the only other feature I could add to Word Owl right now, that I would actually want to, is the ability to search for authors within comics. But even that seems to be non-essential really. To me that suggests I’ve now reached a good place, a place where I develop as it becomes necessary rather than because I figured it was a good idea.

Plus it means I can get the to-do list down!

“The Way Forward” - or “Two Steps Forward Without The One Back As Well”

This is a particularly interesting time - at least it is to me.

I’m at the “Hold the Mayo” stage in development. I have developed everything I think Word Owl needs, although I know full well there are probably plenty of other things I could do. I know I’ve probably thought about them and decided that, actually, I don’t need them right now.

That’s probably a good thing, actually, since it has stopped me bloating up the site with stuff I don’t need to add, just because I thought it might have been a good idea.

Eagle-eyed readers may notice I’ve cut the to-do list. A Facebook app sounds cool, but right now it’s just not worth it. Same with a Firefox toolbar.

Let’s be sensible here: I carry 3,400 strips in 7 comics. Unless the numbers go wacky quicky, neither is even remotely essential. When I have more strips, it might be worthwhile doing. So I’ve removed them from the to-do list, unless enough people ask.

But nothing on there is particularly huge, and really if I sit down long enough I might actually get them implemented/completed anyhow.

Just part of me likes making new features. So if there is anything people want to see, let me know. If enough people tell me they want something, I’ll implement it. (Aren’t you proud? I’m finally building stuff like a grown-up web developer for a change)

I’m a muppet again. I wrote a nice new feature into the site - random strip from a given comic. Tested it on the development server, worked great.

Except I posted the updated copyright page (which had the link) and not the updated random script which actually does the work.

Can’t believe I’d get that one wrong, but I did. Fortunately it’s fixed now, but it has reminded me to rewrite the daily update script - I need to put something in to check whether code has been updated and if so to flag it up that a code patch might be in order. (Code patches get applied after a day or so depending on how big a change it is)

Anyway. It’s fixed now.

I have requested permission for the next comic to be added to Word Owl, and I’ve had one email on the subject, but I am waiting for a further email before I can start work - couple of things to sort out first.

In the meantime I’m going to work on Bruno tonight…

Just upgraded the blog to WordPress 2.6. Assuming the upload has gone to plan, this should be posted successfully - including onto the Word Owl LJ as well.

If not, there will probably be a lot of swearing…

Current Mood: hopeful

I finally got round to sending the email yesterday - the one I meant to send on Sunday…

Anyway, Matt at C&H said I was good to post these pictures, so here they are. Both of the actual pictures here are thumbnails - clicking on them takes you to a more detailed page for the images.

The Explosm banner, at around 2.30pm…

The real treasure for me, though, was the comic. Now, I think I already mentioned that the C&H guys were doing free doodles for those who wandered up to the booth. They were asking for ideas, and I was very specific - “Two guys, running from a giant owl, with the back guy saying, ‘Run! It’s eating everything I say!’” - being a reference back to Word Owl.

And here is the comic. I haven’t been shopping round anywhere that sells frames, but once I’ve managed to find the right frame, I’m going to frame it and put it on the wall in the study if Liz’ll let me.

Anyway, real life has intruded once again which is why not so much else is happening.

I couldn’t think of a proper title for this one, so I won’t bother.

I’ve been wandering round a few forums, looking at what’s going on, maybe doing a little shameless plugging. I have also posted a message in the Bruno forum as I’ve hit a snag with the storyline I have to transcribe next.

I did also write my little comparison of Word Owl and Oh No Robot. I don’t think it’s even that unbiased, actually. Read it for yourself. I’m quite impressed actually. One of the things I mention, and actually put myself down over is the lack of hitting a random strip within a given comic. Before you could hit the Random link and see a strip from any comic, NSFW included. Now from the Copyright page, you can select a random link for each comic - so you can see a random strip from that comic only. I should update the comparison but I want to leave it as it is, as that was the state of play as at this afternoon.

I have also had a look at the comics I might approach next once I’ve got more of Melonpool and Bruno up (as C+H, FFHH and Blank It are maintained as new strips arrive), but I won’t talk about that yet. It’s still a ways off yet, I think.

I’m really pleased with how it’s turning out so far. I did also finally send the email to the C+H guys that I’ve been meaning to all week, which is also cool.

Anyway, that’s enough from me for now.

Current Mood: content

Well, I’ve done the next storyline from Bruno, which was interesting as it is an entire guest week comprising a whole story, although completely non-canon. I’ll add all of today’s update (today’s C+H, Blank It and FFHH, plus this storyline) later on in case I do some other strips today.

I keep meaning to write an article on why I don’t use MySQL in Word Owl, but never seem to get round to it. If nothing else it’s one less vulnerability I have to contend with. I even started making a diagram in Dia but again I need to finish it.

The other thing I need to do is sit down and do the “most unbiased I can” comparison of Word Owl and Oh No Robot. Even putting aside the “what I dislike about it” arguments, that is one of the hardest problems I’ve had to contend with when discussing it with comic authors - and it is a very complex comparison.

Let me sum it up: Word Owl is inspired by Oh No Robot. But has all of the additional features I wanted to add, whilst not implementing things I don’t want to provide; partly because they don’t fit within the sphere of Word Owl, and partly because they would be pointless to maintain based on how Word Owl is maintained.

I do want to make the comparison, but I want it to be as unbiased as I can make it, which is sometimes difficult for me - I’ve spent nearly all of the last 5 months making this. It’s not finished, but hey, what is?

So, expect to see a page somewhere on the site about it soon.

OK, it’s quick update time.

Yesterday, I received permission from John Troutman to include his new work, Flat Feet & High Heels to Word Owl, so I’ve added it to the database. John is very complimentary of Word Owl on his comic, as linked to above - go check it out!! I am highly appreciative of John for this, because it’s a new work, only 8 strips old as at today, so I can keep it up without having a massive archive to include. Now, I can concentrate on getting Bruno and Melonpool up without any other back archives to clear.

In the meantime I have also finally written the script I’ve been meaning to, a one-click updater. (Previously updates were sent to the site as a massive tarball containing all the data, new and old.) While this is slightly longer to run, since it only bzip2’s one part of the update, total bandwidth used is less. Also means that I don’t need to drop the site into maintenance mode to actually perform the update, which is also nice.

I was just doing something with Bruno - flicking back to find a previous strip to check on a character, and I noticed something weird.

The strip was marked NSFW, and yet, I couldn’t work out for the life of me how come. There weren’t any words in there that I would have expected to see make it NSFW, like &#@! or *@!%! so I couldn’t work out what it was.

Then Liz spotted it.

One of the entries on the “this is a naughty word” list is the regular expression \bhell, meaning words that start with hell - you can see how this would normally be used for NSFW strips, but it also meant it was catching words like Hello.

This has, of course, now been fixed, and another 23 strips across the archive are no longer NSFW because of this problem. (But, do I feel st00p1d!!)

Current Mood: amused