WordOwl

Unless the situation changes drastically, there is going to be a delay in updating strips over the next few days.

I am not closing the site; search facilities will continue to be available during this time, however I will have at best limited access to email, and no facility to update strip details, hence the delay in updating.

This is caused by real life situations which cannot be avoided. I am also not taking this opportunity to hand updates over to any third party, although that has been considered. Specifically there is only one other person that I have discussed any part of the technical structure of WordOwl with and thus would have some of the knowledge concerning the update process.

But the reality is that despite the fact that although we have discussed the possibility, it is not appropriate nor practical to burden them with this, and thus updates will have to be on hiatus until I am in a position to continue them.

The current plan is as follows:

• Thursday updates - updated as normal
• Friday updates - updated as early as is practical (any strips released after 9.30am GMT will not be updated on time.)
• Saturday - delayed
• Sunday - delayed
• Monday - delayed
• Tuesday - delayed
• Wednesday - all strips will be caught up to date, probably no earlier than 8pm GMT.

A further update will be posted when this is all up to date. In the meantime, Monday’s “Comic of the Week” article will posted ahead of time and scheduled to switch to ‘posted’ on time.

It is interesting how things are turning out right now.

First up, I’ve been falling behind again on my 50/day thing, but the last couple of days I just haven’t had the hubris to do too much. I did poke the code around though, which always cheers me up a little.

The tag system is now pretty much what the storyline system should have been in the first place, but it was one of the first things I wrote, back in February.

What’s interesting for me is that I’ve effectively removed a feature, although upgrading another; the tag system has been around since July, but all I’ve done to it as add an extra column to one table, add an extra few lines to the DTD and the code to gather the data for it, and about ten lines to actually output it, which I’d already written before for the internal browser (the one that gives me a split-screen view of the original comic and my transcription). It was already there so I could see which tags had been added, but of course the transcription now includes that as standard.

So yeah, I haven’t gained a feature, exactly, I’ve dropped one and extended another, meaning I could rip out a surprising amount of code where I had two different systems doing mostly the same thing.

I’ve also added tags to Sporkman, but that’s only minor really.

I was previously afraid of having to tag every Bruno strip but it occurred to me that I could speed it up a little by using what I already had.

I wrote a small script in PHP to read the XML line by line, if it got to a date to snag the date and work out which tag was appropriate, and when it got to the right point in the block, to output the tag data based on what I already had for continuities.

I still need to do a little cleanup, to re-tag some entries that were previously lumped together, but I’d rather re-tag a couple of dozen strips than a couple of thousand…

Anyway time for bed, said Zebedee.

Right, I have now added the necessary changes to the tag system; tags indicated as being story line tags will link back to the first strip, while non-story tags do not link.

It’s uploading now, via the regular upload process however I think in hindsight I should have done this differently, since the system isn’t smart enough yet to work out which strips need to be re-rendered… all 800 Melonpool strips, all 82 Menage a 3 strips, a goodly chunk of xkcd and a goodly chunk of Cyanide and Happiness are all affected. And it’s uploading each via standard SFTP… what was I thinking?!?

I should’ve tar-bz2 the whole lot, but didn’t think about it.

I haven’t mentioned this much before; I’m not sure if I’ve mentioned it at all, actually.

I’m removing the story arc system as used by Bruno and (formerly) Menage a 3. The system is hideously inefficient, it relied on specifying the start and end dates, and any strips within that date range are considered as within that story arc.

For Ma3 it’s not a problem, but for Bruno there are instances where it is; specifically the most recent arc shows it nicely - the story arc is listed in WordOwl as being 69 strips, but it is clearly 66. The extra three are bonus strips, but as they fall into the right time period.

The new system will deal with that, as it is the one currently in use in Melonpool, so it can easily handle it. By streamlining the two systems into one, I get the benefit of having the flexibility of the tag system and strip out the extra code.

The other benefit I’ve established, although very subtle, is that it does actually provide me a small performance edge. For those who are familiar with Sphinx, the trade off is a Filter against a single MVA list (which is mostly 0), or a FilterRange against many non-distinct values in an integer attribute column. Since my MVA is mostly 0, and even where there are values, they are mostly distinct values. Right now the performance boost is not huge, but every little boost helps.

I’ve already moved the details for Menage a 3 into the tag system, and adjusted it so that the tag details are printed out along with the strip, as the continuity link does currently, although it does not yet point to the first strip (as the continuity link does). The tags are not all storylines, so I can’t link to the first strip in every case, until I add a flag to denote which tags are, or are not, story arcs.

This week, we take a look at another olden goldie, so to speak: Sporkman, by John S. Troutman, the same artist as behind Flat Feet & High Heels.

So, what’s it about?

Sporkman follows the adventures of Sporkman, a superhero whose head is a combination of spoon and fork - thus, Sporkman. It follows his adventures with the Flint family, especially Sophie Flint, with whom in a previous alternate timeline he used to share an apartment with. Now he lives at the Flint family home and has all kinds of naive but hilarious adventures.

What’s the appeal?

Normally, superheros are these suave, sophisticated and above all smug characters with their powers and abilities ruling the day. (Of course this isn’t so true of the modern dark retcons, but we’ll leave those aside for a bit.) Sporkman takes that notion and turns on its head - he’s naive, almost of childlike proportions on occasion - and does things that are cool and mostly right, instead of always being right. Last week, for example, we saw Sporkman turn members of the Flint family into singing and dancing animals, mostly bears, for a jamboree.

In many ways it is the perfect complement for the week to Flat Feet & High Heels - FFHH has darker storylines, although still features the same wit but with a more realistic edge, while Sporkman does not have to conform to the more usual rigours of storytelling, so all kinds of fun and hijinks can ensue.

What about its history, and its future?

Well, Sporkman started around 2000, and as far as is known, has been a member of Keenspot ever since. John decided after a while that he didn’t like the art style of the earliest comics and as such removed them, and in 2004 returned with Sporkman in a newer style, this time chibified. This continued into 2005, but John ended up stopping it to work on his other projects.

In 2008, not too long after the start of Flat Feet & High Heels, John also brought back Sporkman, initially M/W/F but now every weekday, after his fans said they enjoyed Sporkman too.

More recently, there was a time-shift in Sporkman, caused the Nega-Spork, which altered reality so that instead of having two parallel universes (in which Sporkman & the Flint family existed, as well as a second - actual - reality where a slightly different Flint family also existed), John harmonised the two, so that now there is only a single universe, which allows for hints of crossovers between the crazy world of Sporkman and the more subdued world of Flat Feet and High Heels.

Indeed, in today’s Sporkman, there are two characters directly from a FF&HH strip.

Closing thoughts?

Sporkman has come a long way since its 2000 debut. Anyone who either remembers the strip from then, or discovered them via the Internet Archive will no doubt note that the character has evolved, but so too has the artist. It’s been a voyage of evolution and discovery.

As I said, it’s the perfect balance - FF&HH has the darker edge to Sporkman’s lighter edge, FF&HH has more realistic art to Sporkman’s chibi crowd, but the two are now inextricably linked and complement each other beautifully. I’ve actually taken to reading the two directly together (as opposed to whichever order they turn up in my RSS reader) just because it works so well.

Great! Where next?

• Website: www.sporkman.com
• Updates: every weekday
• Created by: John S. Troutman
• My favourite quote: “Well, I told 400 million years of evolution where to stick it.” — Sporkman

Those of you who know how I do things will know that all transcriptions get entered into a custom OpenOffice spreadsheet (useful tabular structure if nothing else) then I press a magic button to build the XML fragments.

Well, I have a custom one for each webcomic, tuned to the specifics of the comic, such as character lists, plus dealing with stuff like the URL (if the URL uses an ID number, such as Cyanide & Happiness, or the InkTank strips, I just enter the ID number and it generates the URL in the strip block for me)

Now, when I first wrote the script back in February I never envisaged half of the enhancements that got made to it, but each is still tailored to the individual comics. It’s now at the stage where I actually make a judgement which one I use as a base when adding a new webcomic.

Sometimes a little more work is needed though, and Melonpool’s is no exception. It was mostly based off the original Bruno template since almost nothing had changed (the space where I added the caption is now where I add the tags) but I finally got fed up going into my text editor and editing in the details of panels without borders. For each panel I’d change it from <panel> to <panel border=”no”> but finally no more. I added a small flag to the transcript as I set it up in my OOo sheet and it works automagically.

This should mean that Melonpool gets done a little faster, but after trudging through my own code I haven’t touched in months (since I haven’t had to, and in any case it’s undocumented!) I finally made the addition.

This was a long post for a small thing, but I’m glad I’ve done it. I just wish I’d done it sooner.

The last set of strips is just being uploaded for xkcd’s transcriptions.

It’s been interesting today. I fixed a minor bug in the parser concerning modified strips (specifically, strips that had no dialogue/other text elements other than a description, title or caption, were flagged as ‘new’ strips because the query figured there was nothing in the textentry table for that strip)

I’ve also finished up, as above, on the xkcd transcriptions. This is the first comic that is properly represented in Oh No Robot, although I can see that my transcriptions are slightly higher quality - there are no gaps in mine (I reckon there are about 5 gaps in ONR, and I can see a couple of strips that have been mistranscribed, e.g. strip 506 is not transcribed at all but strip 505 has been transcribed with strip 506’s URL. There are other examples, but I believe that my transcriptions are slightly higher quality.

And now back to Melonpool and Angst Tech.

Despite spending over an hour participating in a fleet war in Star Pirates (which we lost) I managed to find the time to break out over 100 xkcd strips - this morning (well, yesterday morning, now) I had 360 done, now I have 465 done. By the time the next strip comes out on Monday morning, I should be up to date.

I’ve also gotten the code in for the fix concerning the strips without dates. I haven’t tested it yet against a comic whose strips have no dates at all, but I don’t think I’ve seen one yet.

Anyway, it’s coming up to 4am and I need to go to bed. (I’ve managed today after going to bed at 4am and getting up sometime between 7.30am and 8am and I’ve been in slo-mo all day because of it!)

I’ve been working on transcribing xkcd, and have hit a… situation. It’s not a problem, but one which requires some thought.

Specifically, xkcd ‘404′. It is discussed in the site forums as being an actual comic, the joke being that there isn’t a comic - based off the HTTP code for ‘not found’.

As it is essentially a comic it should be transcribed. The only problem is that it has no date.

Now, previously this would be an issue; the assumption is that all strips have a date, but this is a strip-that-isn’t-a-strip, and as such it has no date.

So I’ve had to jiggle a few things around to deal with it - and with anything similar in the future that is date-less, although it does seem as though this single strip is the only example.

It won’t appear on any of the graphs (since it has no date) and if it does appear in listings because it has matched something else (e.g. ‘not’ or ‘found’) it rank above any other strip of equivalent weight - if you have two strips, this 404 and another strip with the word ‘found’ in exactly once, the 404 would rank higher as it has a date of 0000-00-00. (Standard SQL for none.)

The only slightly annoying issue for me is that I can’t - without an awful lot of work - slide it into the archive page (i.e. selecting xkcd, no search text) at position 404 because it makes no sense to do that - the id number is not intrinsically the order of the strips; it is an arbitrary id number and should only be used in absence of any other defining factor.

But no matter, I have built support in to allow this now, which is a huge step forwards.