Sean Middleditch

Well, after over a decade of dealing with my defective left leg, I’ve finally scheduled an appointment with an orthopedic surgeon to have it fixed.

It hasn’t been more than a nuisance over the years. People can’t even tell there’s anything wrong with, not even under close inspection - I have to point out the problem. The problem isn’t with the way the leg looks, although it does irritate me at times. The problem at this particular moment is the pain its in anytime I put weight on it. The problem in general is the very, very slight angle my ankle is off from where it should be.

Bend you knees. Look down. Your knees will bend outward along a vertical plane aligned with your foot. That’s true for my right leg. Not for my left. I can’t even tell most of the time, until I start really using the leg. Fencing and fighting, dancing, running… they all hurt, and they’re difficult to do well because of the problem I have maintaining balance on my left leg. That angle puts stress in the wrong parts of my foot and ankle that don’t match what the human body was made for.

With all the running and jogging I’ve been doing lately, my left leg has been in near constant pain. I’ve quit the running, but I don’t want to quit it forever. I quit just long enough for the ankle to stop hurting, start running again, and within a day or my ankle is unusable. And it’ll be that way my entire life, possibly getting much worse, unless I get it fixed.

Hopefully the surgeon can do something. I really, really, really don’t want surgery. I don’t want my leg to be unusable for several months like it will be if they do the surgery. But if it can be done, it needs to be done, and it needs to be done soon. I’m sick of not doing things I want to do because my ankle is a few degrees off.

It would have been a lot better to have it taken care of when I was young. A brace as I was growing could have made sure the angle developed properly. Instead, I get to do it the quick but painful way. Assuming it’s even possible. As much as I don’t want the surgery, I’m afraid that what the surgeon will tell me is that it can’t be fixed.

Just bought a new mandolin and acoustic guitar. I paid a total of less than a $100 for the both of them, including shipping. Not bad. Certainly not the best quality instruments in the world, but given that I’m expecting to take these places where water and impact damage are common, not to mention a little too much thievery, I think they’re perfect.

The March KANAR event was a let down. I went Friday evening, only to find that even several hours after the game was supposed to start there was no GM, very little staff, no Treasurer, etc. And only a handful of players. Sure, it was cold and muddy, but the staff needs to be there - that’s their job. I don’t care if they’re volunteers. If they can’t do the job they volunteered for, they should step down. Not even being able to pay my dues for the year due to a lack of a Treasurer, I pretty much just turned around and left.

For Saturday and Sunday, I just really didn’t care enough to drive back out in the rain and cold. So, in other words, I missed my first full event since I started playing. The best part is, I realized that I didn’t really care. Nice to know that I’m not as addicted to KANAR as I thought I was.

Hopefully the April event will be better. Say, maybe the staff members will actually show up, and maybe the weather won’t be trying to drown us in rain and mud. Granted, that’ll probably mean that we’ll be up to our eyeballs in bugs… ::sigh:: Bloody swamp.

As financially stupid as it would be, I’d love to be able to blow a couple million on a real LARP field.

I’m thinking something more like 200+ acres instead of KANAR’s 41, in a land with no swamp but plenty of healthy forest and open grassy areas. Open grassy areas especially. There’s no open space on the current KANAR field at all, just some not-quite-so-tree-filled areas. And swampy mudness. It’d be nice to have a stream running through the land, maybe with a small pond, but only so long as its a nice stream and not just a line of swamp running through the land.

The field should have two towns at opposite ends of the field. Not like the current KANAR ‘towns’, but actual towns with walls and buildings that don’t look like hobo shacks. Towns with an inn in each, a store or two in each, and some general-purpose buildings thrown in as well. The parking lot should be a real lot, there should be a check-in station you have to pass through to get to the field, and there should be some general-purpose buildings out in the areas between the towns, like a farm or two, a guard tower, etc. It might even be nice to put some “ruins” in the least-used parts of the field for atmosphere.

I have no idea what such a place would cost. The land alone would probably be somewhere between $100k and $200k. The buildings themselves don’t have to be too special, but the current craptastic stuff at KANAR cost several thousand dollars a piece, and that doesn’t include all the labor or permits I’d need for the stuff I have in mind. Add in a good sized paved parking lot, the town walls and around two dozen structures, I’m figuring it’ll hit at least half a million dollars. Which is about half a million more than I have for something that would not only never pay back its own construction costs, but would also be a big drain on money. (How many LARPs in one area could there be that could afford reasonable usage fees, enough to cover maintenance and taxes at the very least?)

Still, it would really be nice to have. If I were to put it in Michigan (not sure I would, the weather here sucks the bum) KANAR could rent it once a month, and other groups could rent it on other days. Smaller LARPs could probably run several different games on that land during a single day, since there’s so much space.

Ah well. For the forseeable future, I’m stuck LARPing in a God-forsaken swampy pit of horror filled with barely maintained stick huts for buildings. That combined with the complete lack of faith I have in this year’s KANAR staff doesn’t give me a whole lot of confidence that I’m going to care too much about any of the other KANAR events this year. If I wasn’t obligated to going to the June (and possibly even the July and August) events due to a promise to a friend, I might just have quit. Find something less exasperating to do with my time. Not like KANAR is the only LARP in the area. Assuming I even really want to stick with LARPing at all.

We bought a new printer for the house, an HP PhotoSmart 8400. It’s pretty cool for consumer-level printer: built-in ethernet/JetDirect support and web admin, duplex printing (with an addon we had to purchase), and pretty good quality in both black-n-white and color mode. Only $300 with the duplex option, on sale.

I’m pretty happy with it, so far. Certainly not the best quality printer in the world, but it does everything I wanted it to (that’s pretty much just color and duplex) plus its network configuration was a piece of cake. On screen display gave me a MAC address, configured that in the ol’ DHCP server, restarted the printer and there I was. Set the password on the admin interface and all’s good.

The only real trouble had to do with setting up the thing in Windows XP. First, the driver/software CD seemed to be bad - it kept erroring out during install. Downloading the software installer from HP’s website fixed that issue. The second problem is configured it in Windows when you don’t have/want the full software suite. JetDirect in Windows is bloody stupid. You must create a *local* printer in order to configure a JetDirect networked printer. Additionally, it’s called a RAW network printer, not JetDirect. The HP software itself used Zeroconf (Rendezvous) to locate the printer, which was pretty cool. The pure-Windows configuration method was considerably more complex than configuring the printer in Linux, but the HP auto-detect method was definitely cool. Linux needs to be able to auto-detect network printers and their model/options like that. Pressumably it’s something that needs to be handled in the configuration tool stack and high-level library stack (like libgnomeprint). The Red Hat Fedora printer configuration tool is complete crap (but still easier to use than pure-Windows printer configuration).

Anyways, let the full-color duplex printing begin! (Mainly, manuals and self-written books. ~_^ )

I mentioned the Epiphany bookmarks system in my last blog entry. That got me to thinking about the general flaws of the UNIX file system, and what can be done to fix it (especially in terms of Desktop-VFS).

The Epiphany bookmark system is not like most other browsers. In most browsers, you have a set of folders that you place bookmarks and other folders in. It’s a pure hierarchy. The problems come in here when you have a variety of inter-related folders. Say, for example, I have two folders - one for Work and one for Games. Now, some of the professional work I do is on Games. So should bookmarks related to those projects go in Work/Games or Games/Work? If I decide to put it in both, I must manually file the same bookmark twice. Updating or deleting one bookmark then has no effect on the other. If I decide to only use one, I’ll end up looking in the wrong place half the time for my book. When you have more than two inter-related categories, it becomes truly difficult to keep track of things. I watch my father using FireFox or IE have this problem all the time; he has several hundred bookmarks, and he takes more time trying to find them than he would just typing in the URL. Most users don’t even both filing their bookmarks since the interface is based entirely around filing.

With Epiphany, when creating a bookmark, you have a list of categories. Select the relevant categories and your bookmark is filed. When browsing bookmarks a virtual hierarchy is created based on the categories your bookmarks are in, letting me browse to a bookmarks in the Games and Work categories by going either to Work->Games or Games->Work.

The classic bookmark system is much like a classic file system. Folders and files. The new Epiphany bookmark system is something new, and something. Why should that new system be limited only to bookmarks?

There’s been a lot of talk lately about meta-data based file systems or search utilities. Apple’s Spotlight, GNOME’s Beagle, Microsoft’s WinFS, and so on are all examples. However, these systems are themselves not really the same as a keyword based filing system. Those system work by indexing data and letting you searching on them. You can type in ‘game work’ and any document which appears to be related to games and work (by the computer’s reckoning, based mostly on file name and contents) will be displayed. If I just save a picture I grabbed off the web that is related to some game contract, these systems will not pick it up. Not unless I tag it. In a way that is unique to images, since “tagging” is something that many file formats do not support at all. And the tagging requires me to type in ad-hoc keywords, not easily and quickly selected them from a list. Additionally, these efforts do not deal with the need to replace the file operations themselves, like Nautilus/Explorer/Finder or the various Save/Open dialogs.

My idea is based entirely on the concept of keywords, not searching. A system that automatically selects some default keywords for new documents would be nice, but it isn’t essential. The system would work by replacing the file manager default view with one based on keywords. When you open up your file manager, you’d see a list of categories, which behave a lot like folders - you select one, and then sub-categories or documents are displayed. “Sub-categories” are rather loosely defined here. Actual sub-categories might be useful to support (I’d like to be able to break Work into Contracts and Township, for example, but have that relationship be handled appropriately), but most sub-categories will be like the Epiphany bookmarks system. If I select the Work category, then the system filters out all documents that aren’t in that category. It then examines the categories the remaining documents have other than Work, and displays those as sub-categories.

In addition to categories, general file type (like “Text Document”, “Photo”, etc.) would be selectable as well. So say I want to find that spec on the THOR project add-on I’m contracting for. I might select Work, then Games, then THOR, then Text Document. I would then see a limited number of text documents related to THOR, and I could find the one I’m looking for based on its title. If I have a lot of THOR documents, I might break it into sub-categories. I could select those in any order, because each is visible at each level of browsing.

The browser would get rid of Up in this case, as you’d only move Back and Forward. Back would be identical to Up, as it would always just remove whichever category you selected last. Forward would “reselect” it. Something like the path button bar displayed in the GTK file selector could also be used in the file manager window.

As a side note, the spatial file management idea probably won’t work too well here. Folders are no longer discreet objects, and a deep hierarchy is preferred over a shallow one.

The next task is to fix the Save and Open dialogs. The Open dialog can look a lot like it does not - just replace the folders with categories (no real interface change at all). The file type selection would continue to be useful.

The Save dialog would have the most changes. When saving to the document store (which would be the default, preferably) it would use a very different UI than when saving to a “legacy” location like a remote file share. The UI would essentially be nothing more than a file name entry and a category list. The same as the Add Bookmark dialog in Epiphany. Some extra options for file type and so on might be needed. Using a folder-based view would also work, and will be what most apps will probably use as they would need their dialogs ported (or their toolkit’s dialogs ported).

Implementation wise, this system is pretty easy. It’s really nothing more than a very simple database setup. You have a list of keywords, a list of documents, and a mapping between the keywords and documents. How the documents are actually stored is a bit important; it could be in a special folder, in the database itself, etc. Using a special folder would likely make file access easier and more performant, but has the problem when two files with the same title but different keywords would be hard to store in a sane fashion.

The only really big problem with this system is that its dependent on some application changes in order to work well. For example, when you do a “Save As” in an application, it would only make sense for the same set of keywords as the original document to be used as the default. However, the code that handles a Save As is generally not given a reference to the original document, and the applications around today do not at all keep a set of file-system meta-data around with the document. (This is something I want to fix using the D-VFS API.)

A keyword based system offers a lot more control over document organizing than does a meta-data search system, and is also compatible with such a system to provide the best of both worlds. A hierarchy has numerous problems that have been wrestled with for years now, at least in terms of managing user data.

I wonder how hard it would be to hack up a gnome-vfs plugin to give a category-based filing system a whirl…

A few stories posted on several big geek websites, not to mention some big flame wars on mailing lists, have centered around whether developers of Open Source software should bow to the demands of users.

All other arguments aside, there is a particular reason why OSS developers should not blindly listen to users that I haven’t seen articulated very well yet.

A computer is a tool. A user has a problem, and seeks a solution to that problem. The computer itself is never, ever the actual solution. The computer is a tool that runs software which provides a solution. Furthermore, software itself is not the solution to a problem. It is merely a tool to provide the final solution.

Take accounting software as an example. QuickBooks or GNUGash or what have you are not solutions. The real solution to keeping track of money - or at least the solution in use today - are the several hundred year old accounting practices based on double-entry book keeping. Your software is not that solution. It is just a tool that makes achieving that solution far easier than using a different tool, such as paper and a pencil. The computer certainly isn’t a solution, no more than the table you use to write on is a solution to your accounting problems.

When you’re talking purely about computing problems, seeing these distinctions becomes a bit harder. When you are managing the files on your hard disk, the software is indeed the solution itself. The classic UNIX command line tools are one solution. Nautilus in another. Midnight Commander is one as well. Each are themselves solutions to the problem of finding and manipulating the files on your hard disk.

However, while these pieces of software as a whole are solutions, their individual features are not. Drag and drop is not a solution. It is an interface design element that facilitates selecting one item and operating on it in some way; often that operation is movement or relocation. It is a solution to the problem of “interfacing with the user for moving a file,” but it is not itself a solution to actually moving that file. It’s just an interface element which is itself non-essential. Cut-and-paste works as an interface element that solves the same problem as drag-n-drop. Moving the file is the problem the user has that the software must solve. Interfacing with the user is the problem that the developer has that the interface element must solve.

The difference between those two problems/solutions is at the root of user feature requests being unreliable. Also at the root of the problem is that even ordinary, non-technical users tend to think like a developer - they try to find a series of steps to solve the problem. The steps they choose are often based on past experiene or “common sense.” A user will very often ask a developer to implement some of those steps, never even mentioning the actual problem.

This is where the problem comes in with developers and users conflicting. A user will have a problem and decide upon a solution. The developer may have a different solution to the problem in mind. Due to a combination of communication failures, stubbornness, and even a bit of ignorance, the situation can turn into one where the user perceives the developer as refusing to solve the user’s problem, and the developer believes the user’s needs to be unimportant, because the user and developer are not actually debating about the actual problem the user is facing. Instead, they are debating about a particular solution to that problem.

A good example would be the bookmarks in Epiphany. Unlike many browser, Epiphany does not store bookmarks in a hierarchial folder. Instead, it attaches categories to a bookmark. Many users had a problem where a great deal of bookmarks would be in a single category, and browsing or finding bookmarks was difficult as opening that category displayed all its bookmarks with no further levels of organization. Many users did not communicate this problem at all, though. Instead, they thought about how hierarchial bookmark schemes did not have this problem, and simply said that they wanted a hierarchial bookmark scheme.

However, a hierarchial model is not a solution entirely. Simply storing bookmarks in a hierarchy isn’t a solution to finding bookmarks easier. In fact, you could store bookmarks in a hierarchy and then display them in a flat list. The problem the users had was in the interface, not that underlying model of the bookmark system. The Epiphany developers have found a solution to the interface problem without sacrificing the category system of bookmarks. Epiphany now breaks bookmarks down into sub-categories automatically and dynamically, making it easy to browse large bookmark lists, and also making it easier to create and find bookmarks than a pure hierarchial system would have allowed. The developers found a solution to the actual problem in the interface. If, instead, the developers had listened to the users’ actual feature requests, they would have just implemented a hierarchial bookmark system.

This is the problem with letting users vote on bugs or having weight with feature requests. Users quite often don’t know how to solve their actual problem. They simply try to solve the problem and then expect developers to implement that solution independent of whether the solution is actually any good.

It is important to note that letting users tell developers that there is a problem is still fundamental. You can’t simply take the user’s voice away. Developers need to have the freedom to ignore users’ specific requests and complaints, however, and focus on the real problems. This is difficult at times because finding out what those problems are can be very difficult. Users sometimes don’t even comprehend the real problem, because they are stuck in a thought pattern focused entirely around a particular solution. “Can’t see the forest for the trees,” if you will.

Got a wiki page for Desktop VFS setup on FreeDesktop.org.

Hmm, I actually touched AweMUD and Scriptix today. Just fixed them up a bit to compile with GCC 4.0, which is now in Fedora Core. (A pre-release is, anyways.)

Also submitted my ported gnome-vfs/gssapi patch to Fedora bugzilla (#150132), since they’re more interested in Kerberos than the GNOME project in general. I’m guessing GNOME upstream won’t integrate my patch at all, but will instead just wait until they do the migration to Neon 0.25. Fedora is pretty keen on Kerberos support, though, so I figured they might want to integrate the patch for FC4.

Rampant Mouse is a cool store catering to LARPers. It’s run by the president of the company that runs the LARP I play at (KANAR) and the woman who plays the baroness of the lands in said game. Maybe I’m biased towards the store, then. ;-)

In any event, I promised Todd and Grace that I’d do some work on the site several months back, and tonight just got the first item (but a biggy, in terms of importance) done on the site.

The best part is, it was just a simple configuration error. Sort of. More like a heavily mangled configuration file missing any configuration whatsoever and instead being full of custom code, but hey, that’s what happens when you use osCommerce - that is some of the most opaque code I’ve seen. (That’s not a good thing.) Problem fixed now, yay goodness.

Rum… not bad, actually. Not something I’d want to sit around and drink casually, but now at least I have something to drink in social situations, or something to drink when I’m overly depressed and want to be unconscious. ;-) Which, with my lack of alcohol experience, probably won’t be too hard to achieve…

Too bad I forgot to buy any cider or Coke. Rum and Energy Drinks just aren’t the greatest. Although I’m told Red Bull and vodka go well together… too bad I hate Red Bull and have no interest in vodka.

I’m now patiently awaiting all of the disappoint to be flung my way by the people who were, for whatever reason, happy that i had previously chosen to stay alcohol free, even though being of age for a while now. Ah well - my life, not theirs, don’t care.

Also, for the record, I think in total I’ve not had more than a single shot’s worth. I’m not exactly guzzling it here, people…