User blog:Inept Wiki User/STOP Notes (30/05/2022)

Series Three Overhaul Project - Loose Notes (30/05/2022) These are some loose notes, tasks and instructions concerning the Series Three Overhaul Project (STOP) and other recent stuff on the wiki, as of May 30th 2022.

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

Notices

 * Wiki currently log-in-only due to persistent anonymous vandalism, and will remain so until further notice.
 * Such vandalism will not be tolerated and offenders are liable to be blocked.
 * GT6's arc name has been finalized and the link structure implemented following poll conclusion on the 20th.
 * Check for potential errors with the MediaWiki 1.37 upgrade. Notable CC event late in May.
 * Platform update seems to have occurred on May 24th.
 * Regarding online release of Toaru Kagaku no Mental Out in Comic Newtype, for Chapters 7 onward, only the two most recent releases will be available on the site.


 * Database update requested on May 31st

Recent and Upcoming Releases
Recent Releases:
 * April 8th 2022. Restrictions lifted.
 * Manga Chapters:
 * End of May: ,
 * Mid-May:

Upcoming Releases:
 * Toaru Kagaku no Railgun (light novel): June 10th 2022 (Railgun 15th anniversary project)(Plan thread)
 * Toaru Majutsu no Index: Imaginary Fest: Toaru Senpai no Bridal Collection (Bride event)

Mid-Notes (1)
Due to interruptions, matters requiring urgent attention, and other unfortunate and aggravating circumstances, some notes have been regrettably been skipped.

Slider
During the platform upgrade a few days ago, a number of changes were made. A few things have improved, but as suspected a number of problems have arisen as a result - the most prominent one noticed so far is that the Slider multi-template, used on the wiki's front page, is no longer functioning properly. Clicking on the panel is not switching to the corresponding slides, instead bringing up the image preview, and the slides themselves have lost their intended layout. While wikis with less complex sliders are unaffected, at least one with a similar level of complexity has also gone down. We are keeping an eye on it (alongside many other things), waiting to see if a site-wide patch sorts it out while also looking for potential solutions.

The status of the Slider and Slides below for quick reference: Slider=

Slide 1=

Slide 2=

Slide 3=

Slide 4=

Slide 5=



ImageReqInfo Tracking
Many of the images previously uploaded to the wiki are lacking pieces of information, and a specialized TrackTag known as ImageReqInfo is being used both to visually mark the sections which require information and also track them via Special:WhatLinksHere.

There is also a tracking structure in place which is used to keep track of where and how many IRIs are being used for specific parts of the Summary (e.g. Description, Source, Purpose). The structure which was previously in place involved assigned general TrackTags being placed in the middle of the Description, Source, Purpose etc (e.g. Description: ).

This was derived from a continuation of usage of those TrackTags from when they were inserted into those positions for the complex multi-sweep manipulation operation known as the Image Information Survey, which was executed in September 2020. At the time, many of the older images on the wiki lacked the framework which was being used with the more recent images, and many lacked any information at all.

Rather than manually going through several thousand images, a sequence of tagging and manipulation sweeps were conducted to convert the images' information into the more recent format, as well as marking the current state of those images, en masse and without much individual manual editing work (even so, the operation took about a week of constant work and supervision). The TrackTags were inserted as part of this, and as they carried out the additional function, their usage was continued after the main multi-sweep operation was concluded.

However, over the past couple of years, my colleagues and other have experienced problems with this structure, and have frequently omitted the general TrackTag, included it where not required, or confused it with another set of TrackTags used in the Image Information Survey (one which was used to identify and keep track of cases before that time for further data extrapolation), either forgetting or not knowing. Additionally, there have been some instances where the TrackTags have been retained to mark certain special cases, and records have slipped a bit. I have written multiple explanations and reminders, but these have not achieved the desired effect.

Following a number of recent cases, after thinking it over for a while, on May 20th, I implemented a new tracking structure based upon a series of sub-templates derived from the original ImageReqInfo. Rather than having both the general TrackTag in the middle of Description, Source or Purpose, and the ImageReqInfo TrackTag itself, a single sub-template of ImageReqInfo is used instead. This is functionally identical to ImageReqInfo and keeps track of the specific case where ImageReqInfo is being used - it takes the form, with X being the case it is used in (e.g. for Description,  ).

E.g. Description:

The sub-templates created were as follows:
 * Summary:
 * Description: (WLH: )
 * Source: (WLH: )
 * Purpose: (WLH: )
 * Other bits retain the standard ImageReqInfo (WLH: )
 * Licensing: (WLH: )

While this means that a total list (as opposed to a case-specific list) cannot be acquired from GeneratePageListing or WhatLinksHere for ImageReqInfo, which was possible in the other structure due to the separate tags, additional tools have been acquired, and with them, the GPL/WLH lists for ImageReqInfo and its sub-templates can be combined into one list, with the duplicate items removed, to form the total list. One of the online tools went down recently, but there are others.

There are still items that retain the previous TrackTags, items which did not match the criteria given to MassEdit for the conversion sweeps - THESE ARE TO REMAIN AS THEY ARE FOR NOW, THEY AREN'T FINISHED WITH YET.

Table Notes
Carried forward from previous posts:
 * Possibilities rather than must-haves
 * Only vaguely sorted in terms of priority (a few items have their subject(s) bolded), not strongly sorted at present
 * Subject(s) entries are not image name suggestions. Primarily chosen for descriptive and sorting purposes.
 * No names have been set unless noted.
 * Titles for unnamed characters/abilities likely temporary

Mid-Notes (2)
Due to interruptions, matters requiring urgent attention, and other unfortunate and aggravating circumstances, some notes have been regrettably been skipped this time.

Odd Jobs, Potential Regex Ones
Two odd jobs have been observed or recalled, for which potential regex solutions might be applicable; cases where a letter is directly in front of a pair of square brackets (without a space between them) and cases where a word has been accidentally repeated.

Direct Contact with Opening Square Brackets
The first one is something which is both connected to previous work and which certain users have been making edits related to. It involves circumstances connected to the site's platform change a couple of years ago. Prior to this, text wrapped around images and as such, images were often placed within paragraphs for precise position control. After the change however, images would split any sentence or paragraph they were placed in.

There was something which could be used to locate such instances via sweep - in order not to cause spacing problems in the sentence such images were placed in, they often only had one space on one side of them (two spaces, one on either side, would appear as two spaces in the visible sentence) and as a result, either the opening or closing brackets were in direct contact with part of the sentence, word or otherwise.

For example, as demonstrated in the first paragraph of this section The first one is something which is both connected to previous work and which certain users have been making edits related to. It involves circumstances connected to the site's platform change a couple of years ago. Prior to this, text wrapped around images and as such, images were often placed within paragraphs for precise position control. After the change however, images would split any sentence or paragraph they were placed in.

In the cases where letters were in contact with the closing brackets, there was another sub-project which was focused on cataloguing the wiki's current use of letters after link brackets (rather than file brackets) for plural cases, e.g. witches, in which the search sweeps were inadvertently picking up cases where contact was being made with the closing square brackets for files rather than links.

These were eventually separated from the intended targets and tagged for processing as appropriate (either moving to the top of the paragraph or adjusting/splitting the paragraph). Regex was involved in the initial search sweeps or the separation procedure. However, it was focused on the closing brackets rather than the opening brackets (as that was the area where the two sub-projects overlapped).

In the latter half of May, there were a number of cases where mid-paragraph images were moved to the top of the paragraph (if I recall correctly, there was one for Mugino's Daihaseisai appearance which was moved up after the article was edited and brought up in the RecentChanges display during a reference adjustment sweep), where the opening brackets for a file were in direct contact with words. That reminded me of the previous circumstances, as well as the fact that regex had been involved in the previous operation, and the possibility that regex could be used in identifying the frontal contact cases.

Double Word Cases
Another reminder from observation during operations in late May, was that there were many cases where words had been accidentally repeated in paragraphs, particularly in the older ones (e.g. the the).

A number of AllPage sweeps were conducted using the words that were involved in several observed cases and words imagined as likely possibilities. There was a slight hiccup when the spaces either side were accidentally omitted in one sweep and it accidentally adjusted several cases it shouldn't have done (e.g. for a sweep that missed the spaces either side of "the the", it might affect instances of "the theme" or something similar) before it was aborted and rectified. In addition, there was at least one lyrical romaji case of 'a a' that was affected. However, for the most part the sweeps were successful and corrected a fair number of accidental double word cases.

While the ones conducted so far have caught and corrected many of the accidental repetitions, there are still others lurking around the wiki, mainly in the older paragraphs, and the current approach of picking a likely word and conducting a sweep might not find them particularly quickly or efficiently.

That led me to think about the possibility of a regex formula to locate and tag cases of "_[WORD]_[WORD]_" to narrow down the list and identify the targets (not adjusting them right away in case of potential oversights).

Information Collection
There are quite a few tasks on the backburner involving the collection and assembly of information from the series and various places, but there are a few in particular that should be noted this time around.

One information collection task, partially on the backburner, that has previously been mentioned is finding and assembling a list of instances where fields/branches/disciplines etc of science have been featured/referenced by name within the series, for the table that is to be made on Science.

Following observations made while writing Appearance entries, looking up various things and reviewing past examples, and  have been assigned to an information collection task related to it. At the moment, they merely have an empty table, just the framework. The idea at the time was to assemble existing information and images already on the wiki into a quick-reference guide for stuff that goes into Appearance, thereby minimizing the need to look-up things off-site when writing in the future. SD194 was for stuff relating to physical appearance (e.g. hairstyles etc) and SD195 for attire.

Early June Notes
''Under construction (05/06/2022) Due to the circumstances at the end of May and early June, the completion of this notes post has been somewhat delayed, and certain things will have to be skipped this time around. This has not been helped by an incident at the start of the month that has caused a major depressive attack, as well as a loss of time and progress on the wiki, in the related and previously mentioned communication matter which focus has had to be diverted to, and elsewhere.

However, there are certain things which have occurred during the delayed construction, which will be noted here.

Missing Information from Articles in Category:Light Novel Chapters
Below is an odd job task, following on from a series of check sweeps conducted on June 4th and 5th (making the bot useful while depleted manual attention was being focused elsewhere) for articles in Category:Light Novel Chapters, involving filling in some bits where information is missing:
 * Adapted To sections missing manga details
 * Empty Abilities sections
 * Empty Locations sections

There are a number of related tasks and plans in mind, but to avoid confusion and problems, they are omitted from this post.

Other
A partial discrete overhaul of Level 5 was carried out recently. There are a number of points still to be referenced, or require careful specific referencing, marked in the usual manner by the Cite template.

Preview Override
Series Three Overhaul Project - Loose Notes (30/05/2022) These are some loose notes, tasks and instructions concerning the Series Three Overhaul Project (STOP) and other recent stuff on the wiki, as of May 30th 2022.

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