WordOwl

Yeah, it’s all going kind of swimmingly for a change.

Since I had the HD failure last week and the subsequent recovery, yeah I’ve still lost things. I’ve found a program that should hopefully help me with the final recovery when I get to it (maybe tonight)

I’ve added a ton of comics - since hitting the 5,000 barrier 4 days ago, I’m already pushing towards 5,200. I have also publicly stated elsewhere that one of my favourite comics is one I wouldn’t be able to get on WordOwl because another fan has already made a better one. And it’s sufficiently good that I don’t believe I’d be able to rival it - on that comic’s turf.

Still, where next for WordOwl? Well, I’m pushing towards 80% of Bruno’s archive being complete. I’m having itchy fingers though, there are 3 or 4 other webcomics I’d like to negotiate getting added right now, but I promised myself that I wouldn’t until such times as I had Bruno complete, as that’s the only way I can be fair in my mind to Ian McDonald and Bruno The Bandit - I’ve been pushing it back too much since February when he very kindly agreed to let me add in it.

In my defence, since then I wrote the entirety of WordOwl’s code, transitioned it through removing MySQL dependence, followed by a change to Sphinx, and although I did add all of the other comics, I have been trying to balance it so that Bruno gets some air-time as well as the others, as by adding other comics I get a wider search audience - and thus more people discovering new comics.

So what does the future hold? Well, I still have a few technical articles to document up, such as why/how MySQL was removed, as well as the “How Strips Get Transcribed” I’ve been meaning to do. After that there is the stuff in the to-do list, polishing off some of the new code’s edges, then proof-reading and more comics.

The future’s bright. I just need to make it happen.

On the back of recovering from hardware failure, WordOwl is proud to announce the addition of the 5,000th strip to its archive!

This is not an indication of WordOwl stopping development, simply a point to reflect on that 5,000 strips - and indeed, the first 1,500 official strips from Bruno the Bandit - have been transcribed now, and that more is to come.

We are pleased to announce that WordOwl is back up and running at full strength today after a three day delay due to hardware failure.

Although other data has been lost in the failure, WordOwl’s development files were undamaged so we have run the updates since Wednesday night and are now caught up on everything that has happened over the last three days.

We have also put further measures in place to avoid this happening again in the future.

Current Mood: cheerful

Right, so 24 hours or so on, where are we?

Well, the problem is that my hard-drive has been failing. Now, despite my being a computer tech I had made the mistake of not backing everything up on my home PC (aka development server) on a regular basis.

Much of the really important stuff, like photos and genuinely irreplaceable, is archived on this server (as it’s my rented space) but stuff like the developmental files for WordOwl, plus all sorts of other stuff had not been backed up.

Suffice to say I have learnt my lesson, after spending most of the last 24 hours poking at my drive with a variety of tools and have recovered most of the stuff I can’t easily download. All I’m actually missing is my old “My Documents” folder from when I used to use Win98 (which still housed some of my more recent documents). A chunk of the photo archive lives in there, but all of what wasn’t backed up has been retrieved successfully without problem.

It’s only really little more than odd files and my collection of custom MTG sets, most of which are mine that I’d played around with using the Magic Set Editor, but I am kind of attached to them so I’ll save them if I can. If I can’t, well, I’ll just make them again, only better this time.

In some ways it’s no bad thing, actually, as it’s making me tidy up my hard-drive in a way I would probably never have been motivated to do otherwise. It does also mean that for the immediate few days, probably, no updates at all will be done to WordOwl.

If I say that Sunday is the last day which I don’t do ‘on the day’, assuming all updates do happen as planned, this means I have a backlog of 7 strips (to today), plus 5 (Friday) plus 2 (Saturday) plus 1 (Sunday), leaving me a total of 15 strips, plus the other backlogs to clear.

The weird thing is, the folder in question is one I hardly look into myself, and although yes, I’ll be sad to lose some of the files I’m sure were still in there, maybe I can make a fresh start. Not install any stuff I don’t need, etc. etc.

And in a style unlike me, some song lyrics:

Take a look around you at the world we’ve come to know
Does it seem to be much more than a crazy circus show
But maybe from the madness something beautiful will grow
In a brave new world

Due to a critical hardware failure within the development server, I will not be able to post any new comic transcriptions for a few days until the parts in question are repaired/replaced. Existing transcriptions will continue to be available however; WordOwl.com itself is not affected, merely the development of new transcriptions is.

I am currently working on the relevant repairs to try and resurrect the parts in question, however it may be a few days before I am able to catch up and transcribe comics again.

I apologise for the problems I know this will cause and steps are being taken to resolve this, but as a hobbyist computer user only, I don’t have the funds in order to maintain redundant systems or such like.

Rest assured, as soon as we are able, we will catch up with any backlog that we have.

Although I am happy developing WordOwl, I also realise that there are other languages out there that I’d like to explore. One of the ones I’ve really been meaning to learn is Python.

So, I followed what Randall Munroe of xkcd infamy did (according to his ‘blag’ post, and yeah I know I’m a little behind, I don’t read it regularly), and started trying out the puzzles on Project Euler in Python.

I’ve solved the first three, and in each case I wouldn’t have thought of the solution as readily in PHP, despite having far more experience. I’ve also concluded that I kinda like Python too. If I can find a suitable project to work on, alongside WordOwl, I might try it. (I am considering trying to rewrite some of the guts of WordOwl’s parser magic in Python, but I realise there are major issues around doing this, not least trying to reintegrate with the live version again…)

Although my solutions aren’t optimal, I’m sure, they’re better than what I might have written otherwise. And so far I’m batting 3 for 3 without errors (although I did make a typo entering the answer to number 2 on the site when I did it)

The latest bunch of strip updates has been pretty good; I’ve ploughed through two large-ish story arcs for Bruno in a pretty short space of time, so I’m quite happy about that.

I’m also aiming to do some work on Sphinx itself this afternoon, working on the manual, as well as some house stuff - this week has been pretty tense IRL so I’ve been working on WordOwl to retain my sanity and not put as much time into Sphinx as I would have liked, but this afternoon I’ll switch back to Sphinx, put in a few good hours on adding to the documentation, and drop back to transcriptions tonight.

Then again, that is all dependent on how much of the other stuff like chores I manage to get done too…

I’ve been really bad about posting in here but I haven’t really had a lot to say.

I did a run through of Flat Feet & High Heels in the last couple of days and made a few corrections, plus got all the italics/bold in synch with itself - half was bold/italic and half was bold only. Sorted now though.

The other new feature is a minor usability one: for searches with lots of results pages, instead of: 1 2 3 4 5 6 7 8 9 10 etc, you’ll now see: [1] 2 . . . 9 10 - it displays the first two and last two pages, plus one either side of the current page.

I’m still working on tidying up the new search subsystem to search character names, webcomic titles, and so on, but some other stuff that I’m looking at is making that interesting.

I really need to dig into Bruno’s archive, but a ton of personal stuff has been emerging over the last few weeks so I’ve been working on code; it’s easier to keep my mind active and distracted if I’m playing with code, rather than transcriptions.

In other news, Sporkman has also now gone 5 days a week instead of 3 a week, which is pretty cool.

Changes are coming.

I have a new feature in development, started last night and will hopefully be finished tomorrow or so, whereby searches that match against character names, webcomic names, or storylines and stuff will be displayed, separately to the comic results.

It’s not finished yet, but in the meantime,  here’s a screenshot of what I do have so far. It’s not pretty but it works. Needs a lot of refinement yet though. (Click to make it bigger)

A preview of an upcoming feature in WordOwl's search system

To say that this is beta would be misleading; it’s barely alpha let alone beta. Seems to work as well as Microsoft’s betas of Windows (you know, like Service Pack 2 is still a beta)

I’m also looking at the NSFW system, but more on that later.

Current Mood: hopeful

After early this morning (2am) I punched through the 500 (and 25% of archive) mark with Melonpool, I’ve been digging into Bruno.

Plus my chair arrived this morning so not only am I comfortable, I feel more able to do this.

Aside from today’s strip updates, I’ve also done 55 Bruno strips, and hopefully get a round 100 done today. If I can keep that up, I’ll be up to date with Bruno around Friday next week, assuming 100/day constant rate.

I’m also contemplating a few other things. It may be, for instance, that in a couple of weeks time I have to head off to help with a family affair for about a week, whereupon I may not have daily web access, let alone the development environment to update WordOwl.

This raises 3 distinct options:

1. No updates at all in that time, and I catch up the following week. Not ideal in the slightest, but given the nature of the situation, even if I do get Internet access, it may be better not to do WordOwl.
2. Take the entire development environment with me. I spent a few hours looking at the possibility, and it may just about be doable. The problem with this is the development environment, I need to be able to run OpenOffice.org, PHP (with GD extension), MySQL, WinSCP and PuTTY. Bizarrely I don’t actually need to worry about Sphinx portability; all of the work is done by my PHP scripts outputting a Sphinx-schema XML file and that handles the move.
3. Hand the updates over to a third party for the week. I have a couple of people I can trust to do this, but only one of them has seen the XML data format I use for recording comics, and that doesn’t help with the physical update process, which is nearly all automated now (I update the XML file, run a couple of shell commands and bang, the site is done), but it’s all automated using CLI-mode PHP and WinSCP scripts. Short of asking them to install a Windows VM with the right path layouts and running them through the process, it’s not ideal at all. It’s also not particularly fault-tolerant, but that’s because I didn’t build it to be. I watch the update go and check afterwards that it went, and deal with it if it didn’t go.

As you can see, this is quite troublesome for me, but what I might end up doing is a hybrid of options 1 and 2 - I might try making the entire setup portable, and doing the update that way, but only every 2-3 days.

I’ll have to see how the land lies regarding it and work out how it might be done, or whether I might actually have to just bite the bullet and take a week’s hiatus.

Current Mood: thoughtful