 Userboxes
|
 About Me
My name is Matt. I am an administrator on the English Wikipedia. I aim to make a difference on Wikipedia. Albeit, it may be small, but when I am done here, I hope that I will have made a difference. I am the creator and operator of ListasBot and DefaultsortBot. I am also the creator of WikiBiff. I tend to hang out on freenode as mikaey. |
 About ListasBot
After my failed RfA, another user suggested that I put down the automated tools for a while and learn other aspects of Wikipedia besides just vandal fighting (the area that I was truly active in at the time I ran for adminship). Looking around, I went to WP:BACKLOG to see what I could do to help. When I got there, the item on the list with the biggest backlog was Category:Biography articles without listas parameter, with over 334,000 pages at the time. At first, I said "no way am I going to try to tackle something that huge, I'd never finish it!" But then, after thinking it over for a couple of minutes, I decided to get my hands dirty. After fixing a few of them by hand, I said to myself, "wait, this would go a lot faster if I wrote a program to help me." So, with that, I applied to use WP:AWB, downloaded Microsoft Visual C#, and wrote ListasPlugin -- my first attempt to write anything in C#. Initially, the plugin's function was to give me a list of choices to pick from, and insert my choice into the {{WPBiography}} banner's listas parameter for me. After doing this for a little while, I came to another realization -- some of these pages already had another listas parameter as part of another WikiProject banner, and these pages could be done by a suitably written automated program -- not a huge number, but significant. So, with that, I began taking the ListasPlugin code and moving it to a new project, ListasBotPlugin. I created a new user account for the bot, and applied for bot approval. Having gained approval for the bot, it ran for about a month, save for the break I took in between -- one of AWB's deficiencies, that remains unresolved, is that it starts throwing exceptions at some point after the 32,000 page mark. This meant I had to quit, restart AWB, reload my list, then trim out the portion that the bot had already worked through. I got sick of that for a while, so I suspended running the bot after it made it all the way through the list the first time. At some point during the second run, I was contacted by Magioladitis, who requested that the bot do some more work with {{WPBiography}} banners, since ListasBot was going through pages faster than his own bot, User:Yobot. This became ListasBot 2. One of the functions of this task was to try to determine if the subject of the article was living or not, by seeing if the article was a member of Category:Living people or Category:Possibly living people. At the time, I didn't know about the MediaWiki API, so I relied instead on screen scraping. One of the unintended side effects of screen scraping, however, is that, when the article was a redirect, but the talk page was not, the bot would unknowingly get redirected to another article to screen scrape off of. I discovered this not too long after ListasBot 2 started operation, and this inspired ListasBot 3. ListasBot 3's function was to seek out talk pages that were orphaned when their corresponding articles were turned into redirects, and turn the talk page into a redirect to the new article's talk page. After lengthy discussions with Quadell over exactly how the bot should operate, it was approved and began operation. At about the same time I thought of ListasBot 3, I made a realization that would significantly alter ListasBot's future. It suddenly occurred to me one day that many articles already have a {{DEFAULTSORT}} tag on them, but the talk pages don't. Why not just copy the value of the {{DEFAULTSORT}} tag in the article to {{WPBiography}}'s listas parameter on the talk page? I implemented it, and instantly I saw a drastic increase in the number of pages that ListasBot was able to edit -- instantly, the bot went from editing less than 10% off the pages it was coming across to over 70%. I thought that this sort of a change would have fallen into the realm of "improving the operation of a particular task" in the bot policy, but I couldn't convince myself of that 100%, so I consulted with Jarry1250, the bot approvals group member who approved the original ListasBot. After he responded, I applied and obtained approval for ListasBot 4. At one point, while checking the bot's edits, I realized that, if the bot was flagging the person as living in {{WPBiography}}, it should also be setting the "blp" parameter in {{WikiProjectBannerShell}}s as well. This was implemented as a bugfix in ListasBot 2. However, this gained Magioladitis's attention once again, and he asked that my bot do some more work with {{WikiProjectBannerShell}}s. Not comfortable implementing this as a bugfix, I applied for approval for ListasBot 5. While I was waiting for approval on ListasBot 5, I realized that, with ListasBot 4 having substantially increased the frequency of the edits the bot was making, that AWB was now holding my bot back. One reason was AWB's bot timer -- bots, by policy, are to edit no faster than once every 10 seconds (or 5 seconds, in the case of more important bot functions). AWB's solution was to start counting once an edit was ready, as opposed to counting from the time of the last save. This meant that the bot might edit once every 12-15 seconds, in ideal circumstances. Another reason was that, on my computer, AWB had a tendency to semi-randomly stop processing pages, to where it needed the user to hit the "Start" button again. So, with that in mind, I took the bot's code and moved it over to a new project -- simply, ListasBot -- a standalone application that does not run under AWB, and fixes many of the issues that I saw with running a bot under AWB. While ListasBot 5 was in trials, MSGJ raised concerns with ListasBot 3's behavior of simply wiping pages and replacing them with redirects. Despite my efforts to get input on the matter while ListasBot 3 was going through the approval process, no one had answered me, so I couldn't really say with any definitiveness that my bot was operating off of consensus. I posted an appeal to the masses once again to try to get consensus, but did not get enough input to say for sure that I had consensus on the issue. So, getting a little tired of asking for input and not getting it, I moved all the discussions to a centralized place, and posted the notice at around 140 different locations, including over 130 WikiProject talk pages, the administrator's noticeboard, the bot owner's noticeboard, the village pump for proposals, the village pump for miscellany, and centralized discussions. The issue is still open for discussion. Now, having read all this, you might be asking yourself, "Matt, you put down the automated tools, and you picked them right back up again!" Yes, this is true. However, at the same time, I've learned incredible amounts -- for example, C#, and the bot approvals process. So all in all, I think the whole experience has been a net gain for me. |
 About AfDs
This is directed towards people that see me vote delete in AfDs, and wonder why I rarely vote any other way. Any good publication must have contributors and editors. In the context of this explanation, contributors are people who contribute content to the publication, and editors are people who remove content from it. Note that it is entirely possible for the same person to play both roles. In order for the publication to be of value to society, the contributors and the editors must be in balance with each other. If you have nothing but contributors, then anything goes, and vandalism would run rampant. If you have nothing but editors, then you would have no content whatsoever. However, when the two are in balance, the contributors and the editors are able to gradually create and refine a product to a point where it becomes something greater than the sum of its parts. I choose to take the part of an editor, mostly because I would suck horribly at the contributor part. However, being an editor is not a complete walk in the park. Whenever assessing something for deletion or modification (whether it be an entire article or just one part of it), one must take a sincere, in-depth, unbiased look at the content. For example, when assessing an article for its notability, sometimes it proves necessary to wade through many pages of Google search results to try and find reliable sources. Depending on the article, this can consume quite a bit of time. One cannot just do a search for the term, and rule it out when the first page of results doesn't have anything useful. One must also put aside any personal feelings that they have towards the subject matter or its author. In short, I choose to be an editor because it is what I feel I am better at, but I also feel that there is just as much work involved in being an editor as there is in being a contributor. |
 Awards
|
