User blog:Inept Wiki User/STOP Notes (12/05/2021)

Series Three Overhaul Project - Loose Notes (12/05/2021) These are some loose notes concerning the Series Three Overhaul Project (STOP), as of May 12th 2021.

Due to circumstances at present, both on and off the wiki, these notes will not be particularly well sorted for now.

=Intro= Currently compiling =Assorted Rough Notes= Currently compiling
 * GT4 Plans: Plan thread being used while new volume being processed.
 * New Volume-related Stuff: There are a few things the same and different regarding our release period and pre-start procedures which need to be recorded (as part of the notes on New Release Protocols).
 * Template:SD: Possible quick-link template for linking to Shared Drafts (e.g. Template:Shared Draft 25 to giving SD25, Example to {{SD|208|Example giving Example).
 * Created on May 16th. Seems to be functioning correctly.
 * Profile Image Assumptions: A number of characters have received infobox profile images based on assumptions.
 * Communication and Edit Stress: Existing stress is sometimes aggravated by normal users and troublemakers. Sometimes and recently, a few core contributors have been inadvertently driving me round the twist with certain actions (partially coupled with previous and existing circumstances), but I can't properly communicate/explain what's up, increasing stress even further as a result.

=Mid-Paragraph Image Search=

Regarding Mid-Paragraph Images
There is a problem which arose after the UCP switch-over, the scope of which I wished to assess.

Before the UCP switch, images where placed in the middle of paragraphs rather than at the top, to either get it closer to the text related to it or to resolve positioning problems which would result from placing it in the usual area. Paragraph text would wrap around it accordingly. After the UCP switch, the paragraph text doesn't wrap and the image causes a line-break and thus a spacing problem. This line-break inconveniently splits the paragraph, often where it shouldn't be split (e.g. mid-sentence).

The objective of this sub-project is to locate instances where images have previously been placed within paragraphs, so any problems can be resolved, either by repositioning the image (e.g. to the top of the paragraph) or through another approach. As it would be impractical and time-consuming to locate all of them by manually going through every single article, I've been using auto-sweeps to locate instances, as detailed in the section below.

Example
Here is an example extracted from DA, with the references removed. The second image is the one affected.

Source
Following the incidents at the hospital, Nakimoto Rizou decided to get rid of DA and dispatched Scavenger to deal with them. At the same time, previously hidden information on organization became accessible to Anti-Skill, resulting in DA coming under attack from both groups. With their branches under attack, DA hoped Hishigata's weapon would turn the tables and aimed to get the abducted Misaka 10046 to him at Seiin High School after being told she was necessary for it to work, though Hishigata intended to sever ties with DA as soon as he got the Sister. The Tarantula was deployed to this end, killing many Anti-Skill operatives but ultimately falling foul of Scavenger.

Output
Following the incidents at the hospital, Nakimoto Rizou decided to get rid of DA and dispatched Scavenger to deal with them. At the same time, previously hidden information on organization became accessible to Anti-Skill, resulting in DA coming under attack from both groups. With their branches under attack, DA hoped Hishigata's weapon would turn the tables and aimed to get the abducted Misaka 10046 to him at Seiin High School after being told she was necessary for it to work, though Hishigata intended to sever ties with DA as soon as he got the Sister. The Tarantula was deployed to this end, killing many Anti-Skill operatives but ultimately falling foul of Scavenger.

Locator Sweep Method
As manually going through every single article for cases of images placed within paragraphs would be time-consuming and impractical, I devised auto-sweep methods to locate articles with potential instances and tag instances of the target criteria, which the bot could be set to run automatically while work was being done elsewhere and other things were being focused on.

The devised methods were large themselves, though smaller than a manual check of every single article, a bit crude, and wouldn't completely locate every single instance, but would catch a large percentage of them and shorten the overall time and effort from what it would have been.

If an image is in the middle of a paragraph, it must have stuff close to it, both before and after it. Therefore, it would be possible to use the brackets and the stuff in close proximity the search criteria - just the nearest letter would be sufficient for greatest coverage. Therefore, the search criteria was for the closing brackets with a letter immediately after it with no spaces between the two (if there was a space between the two, the search would also catch the countless non-related instances of linked words followed by a normal word, and given the visible spacing problem which would result from having one both before and after the brackets, at least one letter would typically be in direct contact with one side).
 * The circumstances regarding frontal bracket search criteria have been not fully analyzed yet and no auto-sweeps involving them have been carried out yet.
 * The possibility of brackets being in contact with non-letter characters have been considered but not fully assessed. No auto-sweeps have been carried out involving them yet.
 * The method doesn't cover the uncommon image slideshows which use the gallery tags, but seeing as there is already a gallery location, variant verification and tweaking project on the go at the moment, those can be dealt with in that one rather than this one.

The method for identifying and tagging as many articles and instances as possible was as follows:
 * 1) Perform an AllPage auto-sweep using the pre-2021 article list and targeting direct contact between the closing brackets of the type used for images and a letter, likely the first of the next word in the paragraph.
 * 2) Repeat the above for all the letters in the alphabet, A to Z, covering both upper and lower cases.
 * 3) Log the results and perform any necessary supplementary sweeps.
 * 4) Afterwards, the smaller list can be gone through and tackled manually, by myself or others.

The new article list for 2021 is small enough to go through manually and since it was after the UCP switch-over, the problem would be more noticeable and likely to be avoided.

TT51 Sweep Results

 * 688 articles found with instances of a letter immediately after the closing brackets (e.g. ]]A), for both upper and lower case letters. Instances tagged with the TrackTag between the closing brackets and the letter.

Oversights and Adaptation
As often happens, there were a few oversights which were noticed after the sweeps were initiated and a few modifications to the plan in order to account for them.

One which was spotted fairly soon after the start was that the criteria (letter immediately after square closing brackets) overlapped with cases where links have been included mid-word for convenience (e.g. Espers, Espers). However, as it might be an idea to have a list of these in case they are needed in the future, the sweeps were continued without change in this area, with the plan being to identify and separate these ones later (a few supplementary sweeps have since been carried out to that end (with guessed endings used for search criteria - e.g. es like in witches) but they were rather haphazard (adding notes to the TT51s) and probably didn't cover them all - they might have to be tackled as we go).

One other sweep which was used for identifying mid-paragraph images rather than mid-word links amongst the selection was one targeting instances with full-stops in front of the closing brackets (as might be found in a caption). These instances have had TT51 temporarily switched to TT54.

As mentioned before, the slideshows (with gallery tags) will be tackled with the other sub-project.

Side Note: Have remembered that there are a number of previous user blog posts which have instances of images mid-paragraph. Will have to review those manually.

Tasklists and Procedures
These notes cover stuff to be done in the next stage of this sub-project, to remind and guide myself and others.

Tasklist:
 * Go through the list of tagged articles (WLH, TT51, listed above and on SD21)
 * Find tagged items on each article.
 * Identify if tagged items are mid-paragraph images (as detailed above) or something else (like a link in a word).
 * Sort and record accordingly. Convert non-image instances from TT51 to another TrackTag (considering TT55). Will record list detailing what each is.
 * Once mid-paragraph images have been separated from the other tagged instances, WLH lists for the two TrackTags will be logged to record articles with one sort and/or the other at this point. These will be utilized going forward.
 * Once the lists are properly recorded, proceed with the mid-paragraph images.
 * Go down the list, find the instances on each article.
 * Deal with those instances appropriately, be it repositioning or otherwise. Remove TT51 from instances after dealing with them.
 * Record what was done for each instance and once this has been done for all instances in an article, cross that article off the list (SD212 again?) and continue down the list, repeating the above process until all items have been dealt with.
 * Tackle the items in the other list (not fully decided on how to approach this one yet, to be written).
 * When pre-2021 items all taken care of, go through the list of articles new in 2021 (sweep aid in SD173 or Special:NewPages etc) and locate/resolve any instances manually, while recording what done.

To be considered:
 * Potential frontal contact sweep criteria and methodology.
 * Sweep/task identify links in words (e.g. witches), and possible endings for them (es in that case).
 * Any other potential sweep criteria that could be used to identify other cases of mid-paragraph images that would not be detected using the A-Z contact sweeps detailed above.

IMPORTANT: LOG ACTIONS PROPERLY

=SD208 Variant Parameter/Location Recap= Currently compiling Full sweep result logs have been included in a previous batch of STOP Notes. This section is going to have a compressed version, omitting the individual page results and presenting the different variants for parameters and locations which were found during the sweeps, along with some additional notes regarding related edits made by others during the sweep period. Hopefully, this will be a bit easier for others to read and understand.

Further gallery tweaks are not to be carried out until the core contributors have had a chance to have a proper discussion on the matter.

TT50 Location Checks ( )
{| style="width: 100%; height: 100%; float: left" class="mwCollapsibleAltered mw-collapsible mw-collapsed article-table sortable" border="0" cellpadding="2" cellspacing="1" ! scope="col"|# ! scope="col"|Location ! scope="col"|Successes ! scope="col" class="unsortable" |Notes ===Gallery=== ==Character Art Designs== ==Gallery== ===Character Art Designs=== ==Videos== ===Design Evolution=== ==Concept Art== ==Gallery>== ==Concept art== ==Video==
 * 01
 * 01
 * 81
 * Auto-Sweep
 * 02
 * 02
 * 1
 * Auto-Sweep
 * 03
 * 03
 * 38
 * Auto-Sweep
 * 04
 * 04
 * 5
 * Auto-Sweep
 * 05
 * 05
 * 18
 * Auto-Sweep
 * 06
 * 06
 * 1
 * Auto-Sweep
 * 07
 * 07
 * 1
 * Auto-Sweep
 * 08
 * 08
 * 1
 * Auto-Sweep
 * 09
 * 09
 * 1
 * Auto-Sweep
 * 10
 * 10
 * 3
 * Auto-Sweep
 * 11
 * 11

==Gallery== ===Video=== ==Gallery==  ===Commercials=== ==Gallery==  ===Commercials===  She tries to retrieve the obi from the washing machine, but she is drenched in laundry water instead.
 * 1
 * Auto-Sweep. Success under ===Design Evolution===
 * 12
 * 12
 * 1
 * Auto-Sweep
 * 13
 * 13
 * 1
 * Auto-Sweep
 * 14
 * 14
 * 5
 * Auto-Sweep
 * 15
 * 15
 * 1
 * Auto-Sweep
 * 16
 * 16
 * 1
 * Auto-Sweep
 * 17
 * 17
 * 1
 * Auto-Sweep
 * 18
 * 18

==Gallery== As Haimura's initial design for the sword in the Maiden of Versailles' character design didn't come across too well, he redesigned the sword. Its redesign is based on the image or idea of "holy relics (hair, bones, blood and other parts of past saints) becoming a sword as there are", and Haimura notes that although it is called a holy sword, it has a bit of a grotesque feeling to it. ==Gallery==
 * 1
 * Manual. Toaru Majutsu no Index Light Novel Volume 12/Chapter 2
 * 19
 * 19
 * 1
 * Manual. Durandal
 * 20
 * 20

Character Profiles
==Videos== ===Trailers=== ===Character & Unit Introductions=== ===Trailers/Openings===  ==Character Showcase Videos== ==Gallery==
 * 1
 * Manual. Toaru Majutsu to Kagaku no Ensemble
 * 21
 * 21
 * 1
 * Manual. Toaru Majutsu no Virtual-On (game) - (Gallery gallery previously processed)
 * 22
 * 22
 * 1
 * Manual. Toaru Majutsu no Virtual-On (game) - (Gallery gallery previously processed)
 * 23
 * 23
 * 1
 * Manual. Toaru Majutsu no Virtual-On (game) - (Gallery gallery previously processed)
 * 24
 * 24
 * 1
 * Manual. Toaru Majutsu no Index (Mobile MMO) (sub-section under ==Videos==, same as previously processed ===Commercials===)
 * 25
 * 25
 * 1
 * Manual. Toaru Majutsu no Index (Mobile MMO)/List of Movesets
 * 26
 * 26

Artworks attributed to Kasai Shin
==Gallery==  ==Videos==
 * 1
 * Manual. Kasai Shin
 * 27
 * 27
 * 1
 * Manual. Toaru Kagaku no Railgun - Special
 * 28
 * 28

Trailers
 ===3D Sample===  ==Haimura Kiyotaka's Illustration== 
 * 1
 * Manual. Toaru Kagaku no Railgun: Spectrum Story
 * 29
 * 29
 * 1
 * Manual. Toaru Kagaku no Railgun: Spectrum Story (under Videos)
 * 30
 * 30
 * 1
 * Manual. Eastern Japan Recovery Support Dengeki Charity Project: Send Those Smiles! Comment & Illustration Notepad
 * }

TT44 Parameter Checks ( <gallery )

 * 174-181 - Manual

TT44 Location Checks ( <gallery )
=Other Logs= Currently compiling

Other
=Other= Currently compiling Innocentius ) - might have been usable for isolating redirects which have images, if the category didn't have so many items.
 * Directly copying the contents of a Category page to Source Editor and selecting convert to wiki-text leads to a break tag between the bullet-point and the link when an image is displayed (e.g. *

Navbox Implementation List Method
If one wants to add a navbox to all pages linked in the navbox, here's a method to extract the list for MassEdit:
 * 1) Copy navbox code onto a page, such as an empty Shared Draft
 * 2) Use MassEdit to place two link-breaks ahead of " " and then following "  " - this will isolate the links.
 * 3) Use an external sorter to sort the whole thing into alphabetical order - this will bring all the links together in a distinct area
 * 4) Remove brackets via find/replace function - the list should now be in a form suitable for MassEdit and can be extracted. The page can then be put back to normal.