WordOwl

Well, this last few days has been interesting.

After I got InkTank added, I promised myself I’d work on Bruno, but somehow in the back of my mind I knew I couldn’t do it justice right now, so I’ve been working on Melonpool. And to that end, to actively motivate myself harder than the procrastinating I have been doing, I actually posted on the Melonpool forum.

Part of me feels pretty bad that I’ve not been working on Bruno, especially since I’ve been doing that longer, but I also factor in that it’s around 55% complete, compared to Melonpool’s just-under-20%, and Bruno has a bigger archive. So when I get to 500 with Melonpool, I will stop that and work on Bruno. And I will not rest then until Bruno is completed. I think that’s the only way I can do it in my own mind. I have to be harsh with myself.

In other news, I’ve been doing some work with the guys behind Sphinx; those who know something about me might have recognised my nickname in the Sphinx forums recently. And I’ve been working on developing the manual, getting it up a bit more straight than it was. I spent a long time Friday and Saturday working on it, and I’m going to spend some more time probably tonight working on it.

Back on track, once Bruno’s up to date, I’ll resume work on Melonpool.

But this of course assumes I don’t get derailed with something else yet. But I don’t believe I should. I like this plan. Let’s hope it goes to plan too…

WordOwl is proud to announce the 12th comic to join the line-up: InkTank!

InkTank, started in April 2008, is a semi-autobiographical look at the technology field in California, as well as being a geek, and a dad, with all the troubles of balancing all of those things, and is drawn by Barry T. Smith, creator of Angst Technology and Weak-End Warriors.

InkTank also features Barry’s blog, discussing his other comic-related work, geek and technology humour and the comic updates Monday, Wednesday and Friday.

Today has been spent doing all kinds of funky analysis with the database. Basically I’m trying to do two things.

1. Work out how I might do a “Did you mean…” feature.
2. Work out how to improve search accuracy by folding multiple variant spellings into a single version. For example, a common exclamation or cry from the cast of Melonpool is the word “Aieeya!” Now, I’ve seen variants of “Aieeeya” and “Aieeeeya”, as well as “Aiieeya”, although all are in principle the same word.

As a result, I’ve been studying the 17,000 words in the database to establish about folding. I’d already folded 1,500 up in the previous analysis, and most of the rules were still good from then, but I did some more work on it.

I’m still mulling over the tweaks I mentioned before, about a half-n-half search, but I’m still so unhappy with how that’s looking that I need to step away and work on the underlying code first, so that when I do have the look-n-feel right for that, the code will effortlessly support it.

Plus I’ve been catching up on the forums of Sphinx and all the comics that are listed here; I try and keep up with everything but the last couple of days I’ve been doing other bits and pieces, which is why I haven’t got quite as much done on Melonpool or Bruno as I should have done.

I’ll finish up on the analysis for tonight and spend the rest of the time tonight working on more Melonpool, I think. Then I’ll work on Bruno tomorrow.

I was emailed today by Xavier Xerexes, one of the editors on COMIXtalk, a news and community site for all matters comic related, regarding a short interview on WordOwl.

I’m really quite excited to have been asked to do an interview, as it’s the first time I’ve ever been interviewed for anything, let alone something I’ve made. You can read the full interview on COMIXtalk: How Many Links Does It Take To Get To The Center Of A Webcomic?

As some of you will know, I do keep logs of what is searched. It’s all documented in the Privacy page, exactly what is recorded, etc.

Anyway, boring stuff aside, I can see from the Sphinx log, which is actually more useful, that people have been searching 1977 since the link went up earlier today (many thanks to Byron, 1977’s artist, for putting the link there)

Now, that’s not a bad thing at all, but what it does mean is that I can see a slightly different take on searching to other things. So I’ve realised there is one major thing I need to do, and reasonably soon.

A help page. I think I need to document up what goes on with our magical advanced search page. It is extremely powerful and as easy to use as it is for me, and the few (two) people I distinctly prodded into looking it at it for me, I don’t think everyone else gets it. I keep seeing searches for comic names, and character names. Now, I fixed the earlier one in the last week or so, but I haven’t yet worked out what to do with the second.

Actually thinking about it, I could do what MediaWiki does (and what Wikipedia used to do) and have a split search system. Basically, a search in MW as standard provides two sets of results - matches by title and matches by content.

What I could do is do a split system whereby it looks for matches in character names and comic names first, and then the regular results.

I haven’t got time to explain now but I might make a mock-up of how it would look later and play with it a bit.

Right, Melonpool is now over 300 strips, so with that storyline (one of the funniest in my opinion), the archive is now over 15% complete, which is quite good news. As a side note, that also finishes up all the web strips which appeared in “The Ultimate Melonpool” (TUM1) so I can put away my read-till-it-fell-apart copy and move on to TUM2 for reference purposes.

The current plan is to hit around 700 strips, being 500 more than where I was when I sat back and really looked at it then I’ll really dig into Bruno’s archive again. The problem with both Melonpool and Bruno is that they are pretty dialogue heavy strips. That said, that’s why they appeal to me, and I wish I’d started this project two years ago when I first thought about it (but I wouldn’t have had Sphinx then!!) so the archives would have been smaller.

But even as I think about that, I realise that had I started this in March 2006 when I first thought about it, not only would I not have Sphinx, but a number of the comics searchable here weren’t in existence, and as nice as it is having some big names to work with, I’ve really enjoyed being able to grow the service as the comics grow themselves - Blank It and Flat Feet teamed up with WordOwl when both their respective archives were single-figures, and Menage a 3 and Cosmic Hellcats are both in the 30-40 strip range, so all kinds of flux can still readily happen. I’m not saying that it can’t with others, but strips this young will evolve more than established comics, and I’m really glad that I can grow WordOwl’s services as these comics grow.

I’ve been contemplating some other stuff in the back of my mind recently, that I’d like to talk about briefly.

The first is a “Quote of the Week” featurette, to sit in the footer of the page. Inspired by the quote in the footer of 1977, I thought about doing a feature where I pick a quote from a different webcomic each week and promote that week’s comic.

For example, I might have as one week:

“It’s a shooting star! Ooh, I wish my career wasn’t in the toilet!” — Amanda, from Sporkman, 25 October 2004

However it’d be a bit time-consuming for me, since I’d have to find a great quote each week, although putting it on each page wouldn’t be so difficult once I’d chosen it. I also don’t think it’d be that huge a benefit since a lot of the quotes do kind of require the context in order to be funny.

I’m still contemplating the Wiki I’ve been threatening to do, however I really need to get more strips up before I work on that too much.

I have also been thinking about some miscellaneous seemingly-smaller upgrades to the site:

• The ability to “edit” your search - i.e. from the results page, hit a link back to the advanced search and edit what you’d already entered to tweak it. It’d help with the “back button” problem that I can’t seem to figure out how to fix.
• A “Did you mean…” feature. Actually that isn’t as small as it sounds but it’s definitely something I’ve thought about.

I will probably be tweaking some other bits and pieces very soon; already there have been some tweaks in the HTML for search engine purposes, since not all pages (like the stats) need to go into search engines, so they’ve been excluded, so the search engines can concentrate on the pages that really matter. (I have been thinking about phasing out parts of the stats section, to be honest; I’ve already removed one graph that doesn’t make much sense. If anyone actually wants the graphs, they can be made easily enough)

Anyway, that’s enough for now. Back to the transcriptions!

Current Mood: accomplished

I’ve been thinking about it for a while and I guess it always looked kinda odd that a myperch.org email address (mine) looks after WordOwl’s communications.

So to that end there is now a dedicated WordOwl email address: wordowl@wordowl.com

It’s on the FAQ page, I may even make it a clicky link in the page footer or something yet, I’m not 100% sure.

But yeah, proper WordOwl email address at last.

The last couple of days, I haven’t added too many new strips; I know I should be adding more but I’m having a phase where working on anything other than the latest strips fills with trepidation, so over the last couple of days I’ve toned it down a little and added only the new strips that come out each day.

But that means today I had 5 new strips, plus the one from yesterday afternoon I didn’t get round to; 1977 has a later update schedule than the others - it updates in the middle of the afternoon as far as UK time is concerned, but it’s added this morning.

I’ve been doing some other stuff that is quite different to my usual haunts, which is a refreshing change - I’ve been charting up some cross-stitch patterns for Liz to work on. It’s oddly addictive as it’s kind of like a weird paint program to draw fun shapes in, where you can even draw between the pixels, so to speak.

Anyway, that’s all for now. Might try and do some strips later, maybe getting Melonpool up to ~300 strips, which would be about 15% of the archive.

I was reading through the query log earlier (the one Sphinx makes of the actual queries), and I guess I was amazed at the terms people searched for.

The one thing I did notice, though, and the one thing I thought people wouldn’t do is something they are: they’ve actually searched for a comic’s name in the engine, and actually that wouldn’t have done anything if it wasn’t indexed text in the comic (which in this case it wasn’t)

So now that’s fixed. Bulks out Sphinx a tiny bit, but it’s worth it. Now makes plenty more sense.

OK, so what’s new?

Well, after yesterday’s announcement of Cosmic Hellcats IX joining WordOwl, the guys behind it, wrote a really great announcement about it! This, like all the other posts I’ve seen, makes me feel great to be writing WordOwl, because it’s something I’m passionate about, and it makes me feel good that it’s inspiring others to be passionate too.

In other news I also fixed a bug I’d never been able to produce before but suspected it would happen, if you ever had a search from the front page that would return more than 10 results (i.e. more than one page of “Best from this comic” stuff). And it happened in a different form to what I expected, but that’s fixed now.

I’m pushing Melonpool up a bit - instead of yesterday’s 9.8% being 199 strips, it’s now 11.3% being 230 strips, but I plan to do more yet.

So it’s back to transcriptions for me…