Maps, Directions, and Place Reviews
Broken report to AIV
This report to AIV wasn't formatted properly - it seems that "/r" was inserted there instead of a newline. Svick (talk) 15:19, 26 March 2010 (UTC)
Javanet Connect Exception Video
A bit slow?
Just started experimenting with STiki on my mac. Going well, and nice work. Two comments. One that it looks like there are too many false positives, and also it feels a bit slow although I am on broadband. prashanthns (talk) 11:41, 24 March 2010 (UTC)
My feedback
Hi, I'm trying STiki and it seems like a really interesting idea, but there are two issues I have with it:
- Quite often, I click on one of the three classification buttons and nothing seems to happen. Once I clicked "Vandalism (revert)" and the edit was actually reverted (and the user notified), but the GUI didn't change to another edit and won't change even if I repeatedly click "Pass". Restarting the application fixes the problem, but it very soon surfaces again.
- The tool should use the standard "Month year" headings and warning with appropriate level, not
{{uw-v2}}all the time.
Also I like to watch the talk page I post warnings to, so I would welcome this as an option. (STiki actually does this.)
Thanks for an interesting tool. Svick (talk) 21:53, 22 March 2010 (UTC)
I've been experiencing the above mentioned problem again. This time, I ran STiki from console and got the following messages:
com.mysql.jdbc.CommunicationsException: Communications link failure due to under lying exception: ** BEGIN NESTED EXCEPTION ** java.net.SocketException MESSAGE: Software caused connection abort: socket write error STACKTRACE: === SNIPPED ~100 lines (see history) === (West.andrew.g)
Svick (talk) 22:29, 29 March 2010 (UTC)
Failed to populate metadata object: RID in question is: 352977278 com.mysql.jdbc.exceptions.jdbc4.CommunicationsException: Communications link fai lure The last packet successfully received from the server was 21á793 milliseconds ag o. The last packet sent successfully to the server was 21á234 milliseconds ago. == SNIPPED ~100 lines (see history) == (West.andrew.g)
Failed to populate metadata object: RID in question is: 353066363 com.mysql.jdbc.exceptions.jdbc4.CommunicationsException: Communications link fai lure The last packet successfully received from the server was 21á739 milliseconds ag o. The last packet sent successfully to the server was 21á292 milliseconds ago. == SNIPPED ~100 lines (see history) == (West.andrew.g)
Personal Experinces and some bugs noticed
Ive given STiki about 3 tries so far. Each version seems to improve on the other. I thought it prudent to share my experiences about this tool, as obviously other vandal fighters are considering new tools all the time. The software does exactly what its said to do in that it compliments tools like huggle. Where huggle gives you an up to date as it happens queue, this digs a bit deeper in the history and catches older vandalism (I do not not use twinkle so i will not make any comparisons to that software) that's usually 3 hours old. It often catches vandalsim about 5 minutes old as well. This is very significant. As we have many huggle/twinkle warriors out there but its difficult to catch those that are overlooked in the queue as these tools are used very speedily, so for a vandal fighter who prefers to spend time finding and spotting older embedded vandalism this is an extremely useful tool. That's what I like.
Now what I've noticed that i think needs improvement (at the time of writing i am referring to the march 25th 2010 version); I think there is a small glitch in the formation of reporting users to AIV. I found that my report was sent after the individual was block, it would be nice if there was a way to check that if a user was blocked before filling (minor incidence) Also i think the formatting is a bit skewy when posting at AIV there. What id also like to see is a warning that you are also editing as an IP when you start up. The time of edit would also be useful (is that available anywhere that i may have missed?)The final comment i have regards to a rollback function. A few edits i did reverted the initial edit of the individual but left other incidences of vandalism (since rollback is not featured). At a slower pace this is fine because youd go thorough your edits and this can often happens in huggle anyway with rollback, when people game the system. It would be beneficial i think to incorporate the tool using the rollback much like huggle does. The program is young though. And constantly changing I applaud this program and encourage others to give it a thorough test to. Sorry if this sounds like a blog but i think for a program thats just starting out its nice to share abit of feedback and experiences from other users. Happy editing, and once again well done. Ottawa4ever (talk) 15:11, 25 March 2010 (UTC)
Anonymous users
This can be used without logging in, right? Isn't that potentially hazardous? We could have IP edit wars or 4chan attacks using it. Can this functionality be removed? fetchcomms? 02:57, 26 March 2010 (UTC)
<- And thanks for understanding. I do like the feel of this tool, and hope that it's uses will be expanded (different warning types, etc.) I also like that it's built on Java, which I feel makes it a more lighter interface and easier to use. fetchcomms? 16:02, 26 March 2010 (UTC)
Relationship to Huggle
nor does Huggle
so does Huggle
Just saying... Gurch (talk) 09:22, 17 May 2010 (UTC)
Razi High School edit
As for the Razi High School, Why did you revert my post? Please let me know your reasons for doing that? I had just added the jersey color of the school team! BTW I am a Razi graduate and used to be in Razi High School Varsity team. Good Old Days!--71.62.41.187 (talk) 00:43, 19 July 2010 (UTC)
Suggestion regarding classifications
There are some more possibilities to consider in the "classification" buttons.
- The edit may not be vandalism, but is otherwise deficient and needs to be reverted while preserving good faith.
- The edit may be questionable but in a way that can't be determined solely from a diff - small changes to dates or numeric values are a common example.
- The edit may not be vandalism, and may not be something we want to revert, but is still deficient in such a way as to require further (re)working.
- The edit may be perfectly fine, but in seeing the part of the article given as diff context, there may be something else wrong.
Other than that, looks like a great tool so far, keep up the good work. Triona (talk) 06:45, 7 August 2010 (UTC)
Warning Notices
I don't think the warning given to the user should link back to STiki... It's fine to do it in the edit summary, but putting it in the warning that the user gets will just confuse them, not to mention the fact that it makes this page prone to vandalism, and it could be considered spam. - EdoDodo talk 12:06, 7 August 2010 (UTC)
U IOWA Vandalism detection research
Didn't know if you saw this:
ABSTRACT: This paper proposes an active learning approach using language model statistics to detect Wikipedia vandalism. Wikipedia is a popular and influential collaborative information system. The collaborative nature of authoring, as well as the high visibility of its content, have exposed Wikipedia articles to vandalism. Vandalism is defined as malicious editing intended to compromise the integrity of the content of articles. Extensive manual efforts are being made to combat vandalism and an automated approach to alleviate the laborious process is needed.
This paper builds statistical language models, constructing distributions of words from the revision history of Wikipedia articles. As vandalism often involves the use of unexpected words to draw attention, the fitness (or lack thereof) of a new edit when compared with language models built from previous versions may well indicate that an edit is a vandalism instance. In addition, the paper adopts an active learning model to solve the problem of noisy and incomplete labeling of Wikipedia vandalism. The Wikipedia domain with its revision histories offers a novel context in which to explore the potential of language models in characterizing author intention. As the experimental results presented in the paper demonstrate, these models hold promise for vandalism detection.
- News: http://www.iowahighereducation.com/2010/09/23/ui-researchers-design-tool-to-improve-wikipedia/
- Full Cite: http://portal.acm.org/citation.cfm?id=1772942
Ocaasi (talk) 11:46, 24 September 2010 (UTC)
Error when running
I got the following error when running for the first time:
$ java -jar STiki_2010_07_17.jar Exception in thread "main" java.lang.UnsupportedClassVersionError: Bad version number in .class file at java.lang.ClassLoader.defineClass1(Native Method) at java.lang.ClassLoader.defineClass(ClassLoader.java:620) at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:124) at java.net.URLClassLoader.defineClass(URLClassLoader.java:260) at java.net.URLClassLoader.access$100(URLClassLoader.java:56) at java.net.URLClassLoader$1.run(URLClassLoader.java:195) at java.security.AccessController.doPrivileged(Native Method) at java.net.URLClassLoader.findClass(URLClassLoader.java:188) at java.lang.ClassLoader.loadClass(ClassLoader.java:306) at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:268) at java.lang.ClassLoader.loadClass(ClassLoader.java:251) at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:319)
I'm running Mac OS X 10.4 with JRE version 1.5.0_19.
Still getting an error:
Exception in thread "main" java.lang.NoClassDefFoundError: java/awt/Desktop at executables.stiki_frontend_driver.initialize_button_panel(stiki_frontend_driver.java:457) at executables.stiki_frontend_driver.<init>(stiki_frontend_driver.java:215) at executables.stiki_frontend_driver.main(stiki_frontend_driver.java:177)
(X! · talk) · @705 · 15:54, 5 September 2010 (UTC)
User warning for old edits
I just reverted several cases of serious vandalism that were 10 hours old - way to go STiki! However, STiki refused to warn the users because the edits were too old, even though the user talk pages were empty. I went ahead and did the warnings with Twinkle. It would be great if STiki did warnings like Huggle in actually comparing the time of the last warning on the talk page and assigning the next higher warning level if the reverted edit is newer. --UncleDouggie (talk) 09:06, 3 October 2010 (UTC)
Bug report
I don't think this is vandalism. I don't know how this script works, so I've contacted the user and those here as well.--Ry?lóng (??) 21:11, 8 November 2010 (UTC)
Stiki down?
I'm not getting any revisions. Arthena(talk) 08:57, 26 November 2010 (UTC)
Need rollback
I just reviewed an edit that was obviously vandalism. In the diff, I saw that the old rev also had different vandalism. The history showed that the previous vandalism was added by the same IP on the previous edit. So, I did a rollback from the history page and then clicked "Vandalism" in the classifier to help out the STiki rating algorithm, even though I knew there was no longer an edit for it to undo. It would be nice to be able to go back to previous diffs right in the diff-browser and execute a rollback, just like when using Twinkle. --UncleDouggie (talk) 08:45, 3 October 2010 (UTC)
Failure to revert
I just classified this edit as vandalism and STiki failed to revert it saying that I had been beaten to the revert. However, this was still the latest edit on the page! After this, I did a manual rollback. A contributing factor may have been that I sat on the prior edit I was reviewing for about an hour while I reworked that editor's good faith attempt into something presentable. I then clicked innocent and was immediately shown the Fahd of Saudi Arabia edit. --UncleDouggie (talk) 10:37, 3 October 2010 (UTC)
I just had the same thing happen. In this case, the same edit was made twice, with the first immediately reverted by someone else. Why did STiki not revert the second edit? --UncleDouggie (talk) 11:00, 3 October 2010 (UTC)
The only way out of this was to close and reopen STiki, after which it started doing reverts again. It showed me logged in the whole time. --UncleDouggie (talk) 12:31, 3 October 2010 (UTC)
Minor edits?
I think it would be good if there was an option to mark STiki's edits as minor. By Wikipedia:Minor#When_to_mark_an_edit_as_a_minor_edit, removal of blatant vandalism should be marked as a minor edit. Arthena(talk) 19:55, 6 October 2010 (UTC)
If an article name contains an ampersand ("&"), then the 'Current-Page' and 'Page-Hist' buttons do not function correctly. For instance, for the article "Maram & Aabroo", the mentioned buttons will send you to "Maram" instead. This is a very insignificant bug, but I thought I would mention it for completeness' sake. Arthena(talk) 19:07, 8 October 2010 (UTC)
STiki Release -- Nov. 28 2010
Greetings STiki users, a new version was uploaded tonight! I encourage everyone to download it and begin trying it out. Per the CHANGELOG, some of the bigger changes include:
- Rollback implemented. Rollback is now an option when reverting edits (as opposed to simple "undo"). For users who have the "rollback" permission, this was straightforward to implement. For those who do not, a more expensive "software rollback", that mimics the native permission, was encoded. To this end, a more informative revert-feedback panel now notes both the technique used, and the number of edits undone/RB'ed.
- Per WP:MINOR, all STiki reverts are now marked as `minor'
- The "current page" and "page history" links now handle special characters in article titles appropriately. In particular, this fixes a reported issue with ampersands.
- User's whose last warning was a "4im" of *any* kind (not just vandalism), will now be reported to AIV if reverted by STiki. The post to AIV will reflect the unique nature of this request.
As STiki grows more complex, it becomes increasingly difficult for me to test all the corner cases. I'd appreciate if everyone continues to post their bug reports and feature requests here. Thank you, and happy reverting! West.andrew.g (talk) 03:14, 28 November 2010 (UTC)
Recent performance hiccup fixed
Greeting STiki users. Something funky happened on the STiki server a couple of days ago, and as a result, new edits were not being processed in the back-end. Thus, these edits were not being added to the queue, and STiki users were only seeing old edits that had little probability of being vandalism.
I was alerted to this issue when I noticed that only 1-5% of STiki-viewed edits were being classified as vandalism (in the past several days). Typically, this percentage hovers in the 33-50% range. Rest assured that everything has been fixed, and that performance levels should once again be as expected. Thanks, West.andrew.g (talk) 21:19, 6 December 2010 (UTC)
Bug
Threaded-POST attempt (reversion) failed: java.lang.NullPointerException at java.net.URLEncoder.encode(Unknown Source) at mediawiki_api.api_post.edit_rollback(api_post.java:149) at gui_support.gui_revert_and_warn.run(gui_revert_and_warn.java:173) at java.util.concurrent.Executors$RunnableAdapter.call(Unknown Source) at java.util.concurrent.FutureTask$Sync.innerRun(Unknown Source) at java.util.concurrent.FutureTask.run(Unknown Source) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source ) at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source) at java.lang.Thread.run(Unknown Source)
Caused program to hang ungracefully and then crash. 930913 (Congratulate/Complaints) 13:14, 17 December 2010 (UTC)
Bug 2
Ungracefully crashed; from cmd, salvaged this error message.
at java.lang.reflect.Constructor.newInstance(Unknown Source) at com.mysql.jdbc.Util.handleNewInstance(Util.java:409) at com.mysql.jdbc.ConnectionImpl.getInstance(ConnectionImpl.java:357) at com.mysql.jdbc.NonRegisteringDriver.connect(NonRegisteringDriver.java:285) at java.sql.DriverManager.getConnection(Unknown Source) at java.sql.DriverManager.getConnection(Unknown Source) at core_objects.stiki_db_con.get_con(stiki_db_con.java:55) at executables.stiki_frontend_driver.reset_connection(stiki_frontend_driver.java:289) at gui_support.gui_ping_server.run(gui_ping_server.java:59) at java.util.concurrent.Executors$RunnableAdapter.call(Unknown Source) ... west.andrew.g SNIPPED some 100 lines ... Caused by: java.io.EOFException: Can not read response from server. Expected to read 4 bytes, read 0 bytes before connection was unexpectedly lost. at com.mysql.jdbc.MysqlIO.readFully(MysqlIO.java:2503) at com.mysql.jdbc.MysqlIO.readPacket(MysqlIO.java:600) ... 21 more
930913 (Congratulate/Complaints) 23:36, 19 December 2010 (UTC)
Spacebar keystroke
Two questions related to how the spacebar is functioning in STiki:
1) Spacebar seems to be advancing to a new diff but I don't know if it's using Vandalism/Innocent/Pass. I don't intend it to do any of these, but if it does, which one is it?
2) Spacebar on most browsers is a way to advance a page downward (page down). This feature would be useful in checking diffs, which don't always show up in a single pane (especially using larger font sizes). The desired functionality could be a global spacebar shortcut to advance the page down or one which only worked after clicking on the diff itself. Is something like this possible? Ocaasi (talk) 18:01, 10 December 2010 (UTC)
Dec. 26 -- STiki down for maintenance
Happy holidays, STiki users. STiki will be taken off-line for several hours today for maintenance -- I am aware of this down time. When it comes back online, old versions of the client will not work -- but there will be a new version available for download.
There is at least one big change in the new version which should make the trouble worth it -- MULTIPLE "revision queues." Whereas STiki's metadata-driven scoring system was the *only* one used in previous versions, "Cluebot" and "WikiTrust" now contribute their scores. I'll post the full CHANGELOG here when everything is online and operating. Thanks, West.andrew.g (talk) 15:54, 26 December 2010 (UTC)
STiki Release -- Dec. 26, 2010
The following is the CHANGELOG from the 12/26/2010 release: (note that a change in communication problems will require that users of previous versions update their software)
- QUEUES. Prior to this version, the STiki GUI had only one queue -- an internal one based on edit metadata. (Note: A queue determines which edits are displayed to a user, based on vandalism probabilities. Techniques vary on how one can arrive at a probability). Now, the "Queue" menu offers an explicit choice among several queues. In addition to previous "metadata" queue, there are two other options: (1) "Cluebot-NG" -- those edits whose score was below the threshold required for automatic bot revert (or couldn't be reverted due to 1RR or other rules). (2) "WikiTrust" -- based on the reputation system of Adler et al.. This is a significant change, and more information is available in the STiki doumentation.
- Minor changes to AIV reporting. Both reports and comments now have some additional detail to help the admins at AIV.
- If the 'diff browser' window has focus, the spacebar key will now fire the 'scroll down' action, per a feature request @ WP:STiki.
- Significant changes to communication primitives. All client calls to the STiki database are now handled by stored procedures, as a matter of security. Frontend GUI users should notice no change as a result. All stored procedure calls are logged and can be audited. This change will break previous versions of the client.
- For the backend reputation component (for users and articles), the reversions of the recently approved "Cluebot NG" bot are now included.
- Back-end STiki scores, rather than being on [-inf,inf], now harness the logistic nature of ADTree scoring to normalize scores onto [0,1] (making them probability-esque). This transformation maintains relative ordering so frontend users should notice no change. This transform was also applied to all historical scorings (API accessible).
Thanks, West.andrew.g (talk) 03:42, 29 December 2010 (UTC)
Source of the article : Wikipedia
EmoticonEmoticon