User talk:Humanbot
From Wikipedia, the free encyclopedia
Questions and comments about my bots should go to User talk:Humanbot or User talk:Grammarbot If a bot is going crazy, please leave a note at User talk:R3m0t as well so that I notice sooner |
[edit] Years
Hi, r3m0t. Could you perhaps expand the script of humanbot to include such things as "1960's". I know these need careful checking (perhaps you could leave a note, as for capitals), but they bug me so much! Cheers, smoddy 19:04, 1 Jun 2005 (UTC)
Some people like these, I think. Apostrophe (mark). r3m0t talk 19:24, Jun 1, 2005 (UTC)
-
- Only some odd American magazine... *spit* *spit*. Cheers for that, smoddy 19:37, 1 Jun 2005 (UTC)
- This is actually somewhat controversial. See English_plural#Plurals_of_symbols_and_abbreviations. I believe it's come up on Wikipedia talk: Manual of Style without any consent. Please don't make the speller check for either way of doing this. Deco 01:11, 3 Jun 2005 (UTC)
[edit] cool
Hey there, this is great! I can see myself doing quite a lot of editing this way when I am bored. :) the wub (talk) 22:37, 1 Jun 2005 (UTC)
[edit] "Wanna" > "Want to"
Just tried it, and now it works :p However, I recommend removing "Wanna" as a mis-spelling of "want to". Granted, it can be mis-spelt like that, but too many song/book/film titles include it as wanna and users who don't see this as clearly as me could be corrected all sorts of things and making them factually wrong. For example:
Good Charlotte's "I Just Wanna Live" becomes "I Just Want To Live" (it's always called "I Just Wanna Live").
The lyric from a certain somebody's favourite artists song, Ashlee Simpson's "La La" goes from "You make me wanna la la" to "You make me want to la la" (the lyric is said as wanna).
Of course, users can check, but some will not know whether it is right or wrong and you get too many corrections which are infact doing the opposite ;) Hedley 23:44, 1 Jun 2005 (UTC)
- Then they should press "Don't know" on the left. I made it for a reason, y'know. r3m0t talk 09:43, Jun 2, 2005 (UTC)
I've remove "wanna" from the text replacement list. r3m0t talk 23:14, Jun 2, 2005 (UTC)
[edit] Bot bug
See Image:Humanbotbug.PNG: looks like a lot of false positives for consoles...? – ugen64 00:44, 2 Jun 2005 (UTC)
[edit] removing plusses and other bugs
See: http://en.wikipedia.org/w/index.php?title=Gold&diff=14560003&oldid=14559887
In addition to fixing "artifical", humanbot removed all the '+' characters from the article; I reverted it.
Econrad 02:09, 2 Jun 2005 (UTC)
- Another one [1] Econrad 02:12, 2 Jun 2005 (UTC)
- It munged some dashes here: [2] I reverted the article. Econrad 02:15, 2 Jun 2005 (UTC)
- This one's messed up, replaced the Goenitz article with Gallantmon [3] I had multiple tabs open, which may have contributed. I reverted. Econrad 02:21, 2 Jun 2005 (UTC)
- Last example: [4]
I manually checked all my humanbot edits, and reverted where necessary. Econrad 02:30, 2 Jun 2005 (UTC)
- I suppose it was too intuitive. On Image:Humanbot turnon.png I warned not to use any other tabs... :( r3m0t talk 09:07, Jun 2, 2005 (UTC)
-
- You are correct; I did not heed the tab warning. I'll help with the oops list, and test the new script tonight. Thanks. Econrad 11:49, 2 Jun 2005 (UTC)
- I (and hopefully others) will look through User:Humanbot/oops. Open each article in a new tab, click on "history" on each tab, view the diff of the change on each tab, revert where necessary. Gruelling, but most pages were not adversely affected. r3m0t talk 10:07, Jun 2, 2005 (UTC)
-
- Should we lay claim to sections, to ensure that we don't have people checking the same pages? smoddy 10:16, 2 Jun 2005 (UTC)
-
-
- Update: the links on User:Humanbot/oops link directly to history pages now. :) r3m0t talk 18:40, Jun 2, 2005 (UTC)
-
[edit] Vandelising an article?
I found "%u2019s" several times in School Counseling. The article did not recognize decades such as 1960's. Also, I notice the the issue with "wanna" above. Maybe, you can program the bot to ignore text inside of quotes, but instead maybe post on a discussion page. Also, I'm not so sure my article really needs clean up.
- Moved to Talk:School Counseling
[edit] Turning the bot off
Should anybody want to turn the bot off, it isn't quite possible, but remove the link on the page for starters. That way, people will be unable to start a new humanbot session. r3m0t talk 10:10, Jun 2, 2005 (UTC)
- The "link on the page" will be directly on User:Humanbot/selectproj in future. r3m0t talk 14:17, Jun 14, 2005 (UTC)
[edit] decapitalizing
The bot atuomatically removes capital letters, which is bad for the start of a sentence. Maybe it could detect words starting with a capital? Twinxor t 00:53, 3 Jun 2005 (UTC)
- Yes, that would be nice, but I'm not sure it could really fit into the current implementation. Meanwhile, the extra warning is OK. r3m0t talk 08:40, Jun 3, 2005 (UTC)
[edit] descendent / descendant
All dictionaries I checked permit "descendent" as a valid alternative spelling. OED has examples of this spelling going back to the 16th century. We probably don't need to change all "descendent" to "descendant". Nohat 18:50, 3 Jun 2005 (UTC)
- Descendent is an adjective, descendant a noun. Be careful, but the change is nine times out of ten valid. smoddy 21:12, 3 Jun 2005 (UTC)
-
- OED says
- Johnson gives Descendant n., Descendent adj., and remarks ‘It seems to be established that the substantive should derive the termination from the French, and the adjective from the Latin’. In the n. sense 1, and the related sense 2 of the adj. -ant is now always used; in the other senses of both, -ent is perhaps preferable, but these are either obsolete or so rarely used as to make the distinction one of little practical moment.
- In other words, that distinction is not really valid in contemporary English. Merriam-Webster, for example, gives descendant as the primary spelling for both the noun and the adjective, but doesn't disprefer -ent in any way other than marking it as a variant. It all seems to me to be a matter of style, and not something that we should by default be "fixing". Nohat 21:39, 3 Jun 2005 (UTC)
- OED says
-
-
-
- "in the other senses of both, -ent is perhaps preferable, but these are either obsolete or so rarely used as to make the distinction one of little practical moment" - I interpret that as saying "there are some cases where -ent is used, but these cases are all obscure." – ugen64 03:29, 4 Jun 2005 (UTC)
-
-
-
-
-
-
- That's not what it says at all. It says that there are senses for which -ent may be preferable, but those senses are obscure. The point is not that the -ent spelling is obscure, but that that a distinction between senses based on the spelling is not relevant in contemporary English. A spelling that gets 600,000 Google hits is not obscure [5]. Less common, surely, but not obscure, and certainly not something that we should be wasting our time fixing, when there are plenty of indisputably misspelled words that need fixing. Nohat 04:06, 4 Jun 2005 (UTC)
-
-
-
-
-
-
-
- Also, note that -ant is listed as the variant form in American Heritage Dictionary [6].
- I'm afraid I must agree with Ugen. Although I'd love to standardize spelling, we should not be automatically flagging something which is not clear-cut wrong in most cases. I was chastised once for changing "connexion" to "connection" in many articles, because some parts of Britain still use the old spelling, and editors there were offended. Deco 00:18, 5 Jun 2005 (UTC)
-
-
-
-
-
-
-
-
- You mean you agree with me, right? I'm saying that we shouldn't be automatically fixing "descendent" to "descendant" because both spellings are acceptable.... Nohat 07:32, 5 Jun 2005 (UTC)
-
-
-
-
-
-
-
-
-
- After looking at the evidence (your google search is rather misleading: consistant, agressive, relevent, wierd, and recieved all have significant (over 700,000) numbers of google hits - recieved has almost 2 million hits), I would have to say that "descendent" is undoubtedly the preferred spelling (although "descendant" can also be used) for the adjectival form; however, the noun form, as I have always seen it written, is "descendant." But a variant spelling that gets less hits than "relevent" isn't really a common variant spelling, is it?
-
-
-
-
-
-
-
-
-
- Nonetheless, I agree that it should be removed from the database, I agree: people aren't likely to realize that descendent is a quite acceptable spelling of the adjective, and is apparently a variant for the noun form - we'd have a lot of unnecessary changes. – ugen64 21:57, 6 Jun 2005 (UTC)
-
-
-
-
-
-
-
-
-
-
- It is true those misspellings are common. However, none of them are listed in dictionaries, so they don't make for a very apt comparison to descendent/descendant. Also, you have to consider the relative prominence of the correct spelling.
-
-
-
-
-
-
-
-
-
-
-
-
'correct' spelling Ghits 'incorrect' spelling Ghits ratio descendant 1,750,000 descendant 669,000 2.6 to 1 consistent 69,600,000 consistant 791,000 88 to 1 aggressive 20,500,000 agressive 809,000 25 to 1 relevant 146,000,000 relevent 829,000 176 to 1 weird 20,000,000 wierd 1,200,000 17 to 1
-
-
-
-
-
-
-
-
-
-
-
-
- Here, we see that descendent and descendant are fairly close (2.6 to 1) whereas all the examples you give have a dramatically larger ratio between correct spellings and incorrect spellings. Nohat 22:54, 6 Jun 2005 (UTC)
-
-
-
-
-
[edit] 411?
Any idea why I get an "error 411" when trying to use the system? — THOR 20:01, 3 Jun 2005 (UTC)
[edit] childrens'
I tried the spelling checker last night just to see how it worked, and out of ten checks, two of them were erroneously changing the plural possessive childrens' to the bizarre children's' . --ScottDavis 01:44, 4 Jun 2005 (UTC)
- The nature of the change is silly, yes, and it would not be easy to fix. However, it is correct. You just need to remove the apostrophe after the s. r3m0t talk 12:09, Jun 4, 2005 (UTC)
[edit] Yay!
I tried Humanbot for the first time today. I was a little disappointed it doesn't work in Mozilla (since Firefox's Find functionality won't search edit boxes), but it's still an amazing time-saver and quite well-done. I'm impressed. My only other complaint is to make it not spell-check URLs (although it might be neat to verify that the linked sites still exist automatically). Deco 01:38, 5 Jun 2005 (UTC)
[edit] Links and Images
Great tool! That said, the bot shouldn't suggest spelling error fixes in 1) external http links or 2) image names. --Andy M. 05:24, 5 Jun 2005 (UTC)
[edit] Mozilla
I notice this userpage says it only works with Firefox. Any chance you could get this to work with Mozilla proper? — Gwalla | Talk 07:39, 5 Jun 2005 (UTC)
- Moz isn't used very much any more, is it? Until you see that Greasemonkey supports Mozilla (check this page) I will be unable to support Mozilla. r3m0t talk 10:09, Jun 5, 2005 (UTC)
-
- maybe no more developed, but used, I think, at least as much as Firefox (I'd even dare to suppose overwhelmingly more). Hey you guys, not eveybody reinstalls his system every 3 months! Although I've a 2.5GHz laptop, I still have my home computer from 10 years ago, which runs comfortably and without problem Win98 and RedHat linux (7.1 - or was it 6.1?) with its 32 MB RAM and 300 MHz CPU! And I'm far from wanting to change it! (Even if Firefox is therefore out of reach...)
- To summarize, I'd also like to ask for Moz (if not brower independent) support. Can't there be a (parallel) maybe less comfortable system to perform the same thing? I can easily imagine that a tiny PHP and/or Java Script could do the job. If I knew more about the file formats, I'd write it myself.... — MFH: Talk 18:48, 23 Jun 2005 (UTC)
-
-
-
- You all just confused me horribly. I'm trying to get this thing working, so I downloaded what I know of as "Mozilla Firefox 1.5". Installed Greasemonkey (The version that supports Mozilla Firefox 1.5), and the script. However I get no box to start it. Not like in the screenshots anyway. Am I using the wrong program? Is there a difference between Mozilla, Firefox, and Mozilla Firefox???--Keitaro 19:35, 14 July 2006 (UTC)
-
-
[edit] Edit preview diffs
This tool will be a lot more useful when MediaWiki 1.5 is rolled out on en.wikipedia. If we just auto-preview the diffs instead of the whole article it'll make checking that the changes are correct a lot easier. R3m0t, have you tried making the modifications necessary to make this work on Brion's test wiki test.leuksman.com? Nohat 07:45, 5 Jun 2005 (UTC)
- Even version one (we are now on version two) was ready for it. It already looks for the "Show changes" button, and if it doesn't find one, satisfies itself with "Show preview". (Note: It actually searches by names, so wpDiff and wpPreview.) Be assured this is 1.5 ready, although I haven't tested. :) In fact, I even asked on the development IRC channel about it being backported - but I can't be bothered to try to understand the MediaWiki codebase at this point when my exams are so close and 1.5 is so near to release. r3m0t talk 10:07, Jun 5, 2005 (UTC)
-
- Awesome! Nohat 18:09, 5 Jun 2005 (UTC)
[edit] NOTICE: About improving the accuracy of changes
This is my policy. I will not try to improve the accuracy of changes beyond what it is now. Capital letters at the beginning of sentences will remain ignored; "childrens'" will keep changing into "children's"; and you will still sometimes get several entries in the green box for one change. The changes are meant to be all checked by humans (I always do) and are not necessarily perfect. If the error rate goes down to be so low that some people think it is perfect, some inappropriate changes will ensue.
If some people accept erroneous changes, they remain their responsibilities. If a word has apparently never made a good correction (only one which had to be reversed) I will remove it from the list. r3m0t talk 10:16, Jun 5, 2005 (UTC)
- I agree that users should be responsible for any changes made with Humanbot, but I don't think a low false positive rate will lead to more errors. Although the effect you describe does exist, it would be countered by the fact that there are less opportunities for error overall. I think we should have the best possible system and this isn't an excuse not to fix problems. If you don't have time, that's fine, I'm sure others would be glad to help. Deco 21:22, 5 Jun 2005 (UTC)
- I agree with Deco, I wish it would be more accurate.
Also, I wish it would be easier to compare the text diffs of the page. Whenever I use the bot, I only see the new version of the page, and it's much harder to check the corrections.nevermind, I didn't read above. - Stoph 00:42, 6 Jun 2005 (UTC)
- I agree with Deco, I wish it would be more accurate.
[edit] fo: -> for:
It's changing fo: as in the fo: Wikipedia to for:. User:Luigi30 (Ταλκ) 17:44, 5 Jun 2005 (UTC)
[edit] External links section
I'm not sure if it's being changed or not but on User:Humanbot someone suggests that "External Link" should be changed to "External links" but I thought the guideline is when there's only one link it should be "External link" (although I can't find where it says that) - am I wrong? -- Joolz 09:24, 6 Jun 2005 (UTC)
- It should always be "External links". This style is not really set in stone, but it is referred to as that in several places in the manual of style, and it is convenient, because people will often forget to change it if they add more. smoddy 10:47, 6 Jun 2005 (UTC)
-
- I think I must be thinking of this: "If there is only one link, some editors use the header "External link", but others use "External links" in all cases." which is from Wikipedia:External links, not that it matters much ofcourse! :) -- Joolz 11:04, 6 Jun 2005 (UTC)
[edit] American and British English differences
Have American and British English differences been taken into account? According to the Manual of Style, we should leave as is. I've run into a few, but have just marked them as false positives, but I'm sure there are many that I'm not aware of. Here are some word lists [7] for Brit/Can/American english....not sure how the coding works for humanbot, but it may be helpful. Just a heads up.--Docoga 06:48, 8 Jun 2005 (UTC)
- Anything on the Wikipedia:List of common misspellings with a bullet point (not "commented out") has been searched for. Almost all entries have been included to fix. Because that page was once used for Special:Maintenance, I trust it. If you give me some examples I can remove them. r3m0t talk 11:03, Jun 8, 2005 (UTC)
- licenced is an acceptable variation of licensed in the US, please stop the bot from making this change. Fawcett5 15:27, 11 Jun 2005 (UTC)
[edit] error in correction
Hello.
Someone just corrected the word "amoung" to "among" on the page minority languages of Sweden. The correct word was, in that case, "amount". Just a notice.
--Fred-Chess 20:05, 8 Jun 2005 (UTC)
- A slip like that happens every so often, hence the bolded text on User:Humanbot: "It automatically sends you to the right page, makes the change for you, even presses the preview button so you can check it! You just need to check the edit is reasonable." Sango¹²³ 23:15, Jun 8, 2005 (UTC)
It looks like you broke a link on The Lord of the Rings page by wrongly changing "teh" to "the" when it was suppose to be the misspelling. You really should check what you are doing before blindly correcting. Sending a user to the preview page so they can check it is a great idea, just make sure that your users are actually taking the time to check. Arctic.gnome 16:28, 10 Jun 2005 (UTC)
- That wasn't me. User:Roger McCoy is the one who should have checked it. I feel I have emphasised enough by now that people are meant to check their edits. I can't really do anything about it. r3m0t talk 14:13, Jun 14, 2005 (UTC)
[edit] Minnesota Commercial
I've corrected a few of your corrections on the spelling of "commercial" to recapitalize the word when it is part of the phrase "Minnesota Commercial Railroad". This is a proper noun as it is the name of a short line railroad (the company's official website is still under construction) operating in Minneapolis, Minnesota. When you check your corrections, please read the sentence and ensure that words that should be capitalized remain capitalized. AdThanksVance! slambo 23:21, Jun 8, 2005 (UTC)
- If that was me, you should have written at User talk:R3m0t. If not, to the other user's page. In any case, you will be able to write at User:Humanbot/selectproj in future about these false positives. r3m0t talk 14:15, Jun 14, 2005 (UTC)
[edit] Completed!
Thanks User:r3m0t, that script is really clever and very useful, now that it is completed, what next? Has every article been checked now? if so i guess it should be re-run in a few months or something, or are there similar scripts that could be run to correct other kinds of errors? thanks again. Bluemoose 15:31, 9 Jun 2005 (UTC)
- Just look at my todo list at User:R3m0t. r3m0t talk 22:40, Jun 9, 2005 (UTC)
[edit] Just an idea for Humanbot...
Could you say on the User:Humanbot page whether or not there are open projects? Because I just spent the last few hours trying to get FF to work with Greasemonkey (I was using a nightly that GM didn't like), and after all that, all edits are done. Just a suggestion. ral315 21:23, Jun 9, 2005 (UTC)
- Don't worry, there will be new stuff soon. Anyway, I have put up such a notice. r3m0t talk 22:36, Jun 9, 2005 (UTC)
[edit] Thread mode articles
I believe your bot should avoid thread mode articles like Wikipedia:WikiProject Stub sorting/Criteria, since it's not polite to edit other users' comments. --cesarb 01:21, 11 Jun 2005 (UTC)
- Also Wikipedia:General complaints. And Wikipedia:Unusual requests. And possibly others I didn't see. --cesarb 20:47, 11 Jun 2005 (UTC)
-
- And now Wikipedia:Bot requests and Wikipedia:Disabling edits by unregistered users and stricter registration requirement. --cesarb 02:32, 12 Jun 2005 (UTC)
-
-
-
- Perhaps. It has done good on the document mode pages on the Wikipedia: namespace, but not on the thread mode pages. The problem is that, unlike the talk namespaces or the article namespace, the Wikipedia namespace is a mix of document mode and thread mode, and it's harder for a program to distinguish between both. I'd expect the fact that Humanbot needs human confirmation would avoid that kind of problem, but looks like it didn't. --cesarb 18:04, 12 Jun 2005 (UTC)
-
-
[edit] List of articles with errors?
Is it possible to create a list of articles where there have been errors found? One thing that I've noticed with this project, as well as with other projects on the Wikipedia that have looked for specific spelling errors, is that when then find one or more specific errors, I can usually go through the same page using SpellBound and find three or four more spelling errors. What I usually do is when a page on my Watchlist has been corrected for specific spellings is that I will then go through and do a general spell-check. BlankVerse ∅ 15:09, 11 Jun 2005 (UTC)
[edit] Humanbot
Hi, I'm from the Spanish Wikipedia. I'm really interested in Humanbot, is there anyway I could install this bot over there?--Fito 20:28, Jun 14, 2005 (UTC)
- I will have to host it, because it isn't too difficult for it to become some sort of mad auto-blanker. ;) I will need a list of common misspellings, similar to Wikipedia:List of common misspellings. Also, please translate the relevant messages, including "on", "off", "home", "auto-edit", section 0 of User:Humanbot, section 4, User:Humanbot/takeproj, and others. r3m0t talk 07:55, Jun 20, 2005 (UTC)
- Thanks, I'll get to work on that list later today. By section 0, do you mean the "special script"?--Fito 16:48, Jun 20, 2005 (UTC)
Okay, I created es:Wikipedia:Lista de errores ortográficos comunes, it's a small list, but it will grow. The translation of the script is located at es:Usuario:Humanbot/script. The location of the bot is es:Usuario:Humanbot and here's es:Usuario:Humanbot/takeproj. I'll translate the instructions as soon as we get the bot to work so I can take the screenshots. Please tell me if something's wrong or missing. Thanks a million! :) --Fito 01:07, Jun 21, 2005 (UTC)
[edit] New projects?
It seems I arrived here a bit late. Are there any plans for starting a new project? The idea behind this bot (and the implementation -- mmm, Greasemonkey) look great; I just wish I could try it out! :) -- Nathanael Barbettini June 29, 2005 04:37 (UTC)
I got here when someone suggested it for correcting a large-scale misspelling that is just one of many on the page that lists Wikipedia misspellings. I downloaded the greasemonkey and user script, but it looks like I need to wait for a new project? I'm confused as to how this all works, but it looks very useful and educational. I'll put a follow-up on assistance. Thanks. Spalding July 4, 2005 19:29 (UTC)
Both of you must wait. The next run is unlikely to be a spelling one - it takes 90 seconds to find a word in a wikipedia dump, and few mistakes have been introduced since the last runs. I still hope to make something, though. r3m0t talk 22:10, July 24, 2005 (UTC)
I too look forward to a new project. 71.112.105.18 03:02, 17 October 2005 (UTC)
[edit] Greasemonkey Bug
A story has appeared at Slashdot saying that there is a serious exploit bug with Greasemonkey allowing arbitrary file execution. It is recommended that Greasemonkey be uninstalled as the exploit bug is so serious. This is, naturally, a serious problem for Humanbot should any new projects come up in the near future. David Newton 16:59, 19 July 2005 (UTC)
- They fixed it. http://greasemonkey.mozdev.org/ — Ambush Commander(Talk) 17:19, July 19, 2005 (UTC)
- Unless Humanbot requires the special GM functions. — Ambush Commander(Talk) 17:29, July 19, 2005 (UTC)
[edit] It doesn't work at the moment for some reason. Can you help?
I have installed the User script and it shows up in the "manage user scripts" window. The problem is that the little box on the left hand side of the screen (between "toolbox" and "other languages"). Its nowhere to be found. I've tried a restart and reinstalling and purge but it won't appear. Is there a step that I've missed? Please help... --Celestianpower hab 16:43, 18 September 2005 (UTC)
- I have FireFox 1.5 RC 1 with the compatible greasemonkey, but it is providing a "Component is not available" error in the Javascript console and failing to run. Is this a limited problem or has the script not been updated for FF 1.5 yet? --MattWright (talk) 10:36, 10 November 2005 (UTC)
-
-
- The box should appear, but the bot is no longer operational. You won't be able to do much. Deco 03:11, 29 November 2005 (UTC)
-
-
-
-
- Sorry for imposing, but why isn't operational? I would think this bot would be of much help to the many users involved in the typo project.
-
-
-
-
-
-
- Confirmed. Just downloaded everything and the box isnt there. I was really hoping to try it out. American Patriot 1776 03:57, 28 December 2005 (UTC)
-
-
-
Sorry for not noticing this. The problem is that "As of Greasemonkey 0.6.4, however, user scripts now have their own JavaScript context and execute completely separately from the content document. This means that except by noticing changes the changes which they make to the DOM, content scripts can neither detect Greasemonkey scripts, nor interfere with them." - in other words, 0.6.4 broke compatibility with some scripts in order to secure Greasemonkey. This is a moot point anyway, considering that there are no active projects. I am the only person able to add new projects and I currently don't have a way to collect the data I need.
I have work to avoid today, so I may try to fix the script a little. Then the trouble will just be getting a project. r3m0t talk 14:53, 29 December 2005 (UTC)
[edit] When will he be here?
When will he be back up? —The preceding unsigned comment was added by Firefoxman (talk • contribs) 22:08, 14 December 2006 (UTC).