User talk:Kndimov/Archive 2

I happened to notice that you recently archived a few talk pages. I also do this when I run across pages with significant quantities of very old sections. However, I find it easier to let either User:lowercase sigmabot III (lcSB3) or User:ClueBot III (CB3) perform the actual archiving. For a few reasons, I prefer using lcSB3. I have a couple of standard direct paste in blocks of text which make it very easy to set up archiving and just leave it to be archived by the lcSB3 within the next 24 hours. It is necessary to come back and verify the archiving after the first bot run due to various date issues (missing signatures, etc.) and some instances of malformed text in some situations. However, that can almost completely be handled with a few clicks using User:Equazcion/OneClickArchiver. I generally choose to set the archiving at 2 years, 1 year, or 90 days depending on the activity level on the page. Sometimes I choose some different amount of time due to how the discussions on the page happen to be grouped.

The paste-in blocks of text I use are:

<!--lcΣB3 90 days:-->
{{Archives |auto=yes |search=yes |title=[[Help:Archiving a talk page|Archives]] ([[{{#titleparts:{{TALKPAGENAME}}|1}}/Archive index|index]]) 
|bot=lowercase sigmabot III |age=90 |style=margin-top:2px; }}<!-- The length of time in age= and units= 
parameters in the "Archives" template (above) should match the length of time in the "algo=" parameter within the MiszaBot config (below).
Note that the "algo=" parameter is in days and the "age=" and "units=" parameters are in whatever units are in "units=".
The "algo=" parameter is what specifies the length of time for which the bot actually looks.
The "Archives" template specifies only what is displayed in the archive box.-->
{{User:MiszaBot/config
|algo=old(90d)
|archive={{subst:FULLPAGENAME}}/Archive %(counter)d
|counter=1
|maxarchivesize=100K
|archiveheader={{Automatic archive navigator}}
|minthreadsleft=4
|minthreadstoarchive=1
}}
{{User:HBC Archive Indexerbot/OptIn
|target={{subst:FULLPAGENAME}}/Archive index
|mask={{subst:FULLPAGENAME}}/Archive <#>
|leading_zeros=0
|indexhere=yes
}}__TOC__{{clear}}

or

<!--lcΣB3 2 years:-->
{{Archives |auto=yes |search=yes |title=[[Help:Archiving a talk page|Archives]] ([[{{#titleparts:{{TALKPAGENAME}}|1}}/Archive index|index]]) 
|bot=lowercase sigmabot III |age=2|units=years |style=margin-top:2px; }}<!-- The length of time in age= and units= 
parameters in the "Archives" template (above) should match the length of time in the "algo=" parameter within the MiszaBot config (below).
Note that the "algo=" parameter is in days and the "age=" and "units=" parameters are in whatever units are in "units=".
The "algo=" parameter is what specifies the length of time for which the bot actually looks.
The "Archives" template specifies only what is displayed in the archive box.-->
{{User:MiszaBot/config
|algo=old(730d)
|archive={{subst:FULLPAGENAME}}/Archive %(counter)d
|counter=1
|maxarchivesize=100K
|archiveheader={{Automatic archive navigator}}
|minthreadsleft=4
|minthreadstoarchive=1
}}
{{User:HBC Archive Indexerbot/OptIn
|target={{subst:FULLPAGENAME}}/Archive index
|mask={{subst:FULLPAGENAME}}/Archive <#>
|leading_zeros=0
|indexhere=yes
}}__TOC__{{clear}}

These are included here because you might find it easier in the future to use something similar rather than manually archiving pages. In order to complete the setup of the archive indexing I then preview the page and open the red "index" link in the archive box in a new tab. The Archive index page then needs the text <!-- Legobot can blank this --> placed on it and saved. I then save the main talk page. — Makyen (talk) 03:22, 28 March 2014 (UTC); add comment to both explaining to future editors relationship of time parameters. 03:46, 31 March 2014 (UTC)

Thank you for the helpful tips. I read the page on archiving and the bots, but I figured it wouldn't be too hard to do it manually, considering how most of these pages have not been archived in over five years. Because of the little traffic I didn't think it was necessary to bother with a bot. -- Kndimov (talk) 13:18, 29 March 2014 (UTC)

I find using the bot easier, less time consuming and it solves the problem for an extended period of time instead of a one-off. I have the above text as two separate permanent entries in my clipboard. Dropping one in is just a couple of key presses. Setting up the archive indexer bot is just another permanent clipboard entry and some mouse clicks. Determining how long to set the dwell time to is scanning the page to see the level of usage and if there are usage clumps. Setting up a bot solves the problem for the foreseeable future instead of expending what to me is more effort to solve it only currently. Setting up the bot also allows for people to get used to archiving happening on a regular schedule instead of at what to them appears a random schedule when done manually. For the permanent clipboard, I use ClipX, but there are a good number of different programs available that have similar functionality available on multiple platforms. — Makyen (talk) 23:15, 29 March 2014 (UTC)

I see your point. The only problem that might arise would be if there is a sudden burst of postings on these otherwise deserted pages. Then the timeframe of the bot would have to be shortened, and then changed back when things cool off. -- Kndimov (talk) 13:45, 30 March 2014 (UTC)

I agree that the dwell time for messages on the page should be adjusted based on the activity level on the page. I look at it as the responsibility of the editors actively contributing to the page to keep up with any needed changes in this area. My experience is that once the templates are on the page, other editors are much more comfortable making changes to the |age= and |algo= parameters, or other parameters, to make the archiving more appropriate for their needs as they change over time. This is one of the reasons I use lcSB3 instead of CB3. CB3 requires the age to specified in hours, which is much more confusing to uninitiated editors. I look at it as I am just putting in the framework to help establish maintainability on the page. I intentionally select dwell times that are longer than I might choose if I was actively participating on the page so that it is less likely the bot will initially disrupt any ongoing activity. I keep the page on my watchlist at least through the first bot run, often much longer than that. If there is a complex section which will be archived in the future, I watch to make sure the section does get archived properly. Sometimes this is months later. The dwell time I usually select is often a trade-off between archiving old, inactive threads and leaving active ones on the page. I also try to make it such that the first run of the bot does not archive any moderately recent RfC, closed move discussion, etc., so that people have time to get used to the fact that archiving is occurring. Obviously, if such section is years old, that is not an issue. — Makyen (talk) 20:39, 30 March 2014 (UTC); correct omission of |algo= 03:46, 31 March 2014 (UTC)

Thank you for your suggestions. I have set up an archive bot and will wait until tomorrow to make sure everything is ok. -- Kndimov (talk) 21:05, 30 March 2014 (UTC)

I assume that you are talking about Help talk:User contributions. Unfortunately, that is a bad example. It needs manual cleanup prior to a bot working on it due mostly to the extensive use of level 3 1 section headers. Because the bot is going to start its run today in a couple of minutes, I am going to disable the bot configuration on that page for the moment. More in a bit... — Makyen (talk) 22:58, 30 March 2014 (UTC); correct header level 23:04, 30 March 2014 (UTC)

I figured something was different about all this. So you just separate the level 1 headers across multiple archives? Or are the subdivisions going to be merged after each round of archiving? -- Kndimov (talk) 23:06, 30 March 2014 (UTC)

The use of level 1 headers is unusual for a talk page. I'm not sure how the bot would have handled it. With level 3 (L3) headers it assumes all L3 under a single L2 count as a single section and there must be no date in anything under the L2 which is newer than the cut-off in order to archive the L2. The entire L2 is archived as a single section including all L3 under it. For L1, I am not sure. It would probably treat them the same way that L2 and L3 are treated (all dates under the L1 must meet the criteria for the L1 to be archived). It might treat the L2 sections separately. L2 is standard for talk pages, and the bot may not be programed to consider L1 (I would have to check the source code, or ask the programmer). If it did treat the L2 sections separately, then the sections would become disconnected from the information as to where they came from.

One of the significant goals of archiving is to preserve attribution information. This includes who wrote something, when and where. In this case the L1 headers were supplying the information as to where the text was originally written. The sections should not become disconnected from this information when archived. It is one of the reasons we are supposed to include the page name of the archive in the edit summary if we manually archive something (I'm not perfect about doing so, but I try). [Note: For semi-manual archiving, User:Equazcion/OneClickArchiver helps in this regard (it was not used in this process as none of it was suitable for anything other than manual handling).] As a result, one of the goals of the changes I made was to preserve this information both in the manual archives and set it up such that the information will be preserved by the bot when additional sections are archived.

So what I did:

• Disable the bot temporarily. It begins its run at 0 UTC which was 20 minutes ago. I did not know if I would complete these changes prior to it starting. I did not want the bot to come through making changes while I was partially done.
• Check for and move any (1) sections which had been erroneously placed within a L1 section which was inappropriate. To do this I did a history revision search to find when those L1 sections were added and what was added at that time. One section was added on the page between those two L1 sections being placed on the page. It was erroneously in the first L1 section when it should have been in "Wikipedia" L1.
• Check for activity in the first L1 section. There was no recent activity.
• Move entirety of the first L1 to the archive.
• Check for activity in the second L1 section. There was recent activity.
• Move the L2 sections of the second L1 which have no recent activity to the archive. (I defined recent activity here as 1 year, which is a bit shorter than the bot, but it appeared that the discussion was complete.)
• Increase the header depth L1->L2 L2-L3 for the entirety of the remaining second (now first) L1. This puts the remaining sections which were moved from Wikipedia talk:User contributions under one L2 which will all be archived by the bot at the same time to preserve the information of where they came from.
• Remove the remaining L1 header as it was no longer needed to differentiate the sections.
• Re-enable the bot. Changing the |counter=2. Because I did not change the existence of L1 headers in Archive 1, any additional archiving needs to be on a separate page or it would be erroneously contained in the last L1 on the Archive 1 page. I could change the L1->L2 and L2->L3 in the archive, but the spirit of the archive is to be as close as possible as what was on the talk page. Thus, I felt it better to leave the archive with L1 headers. Another goal is, of course, to disturb the formatting of the current talk page as little as possible (i.e. assume that people want it basically the way it is, even though that is not an accurate assumption much of the time).

That's what I did. It was an unusual situation.

I wanted to mention a couple of things about the process you followed for putting the bot on the page. It looked good. However, there were two minor things. You created Help talk:User contributions/Archive 1 with a header. There is no need to create that page. The bot will create that page when it runs. It does not hurt to make that page, but it does not help either. You were probably trying to follow what I had said about the archive indexer bot and needing to put <!-- Legobot can blank this --> on a page. The page where that is needed is Help talk:User contributions/Archive index. It is the index link in the archive box. I have not made the page. I figured you would want to do so. There is no hurry on creating it. The archive indexing task of legobot is having issues. It is unlikely to run in the near future. — Makyen (talk) 00:41, 31 March 2014 (UTC)

Now that lcSB3 has run on the page, I took a look to see if there were any issues (e.g. section which were not archived which should have been, sections without dates/signatures, etc.). For sections without signatures, I use history revision search to determine when and by who the material was added. I then add signatures using {{subst:unsignedIP2| (UTC)|}} or {{subst:unsigned2| (UTC)|}} if it is not an IP. The " (UTC)" is needed in the first parameter in order for any of the bots/scripts to recognize the date.

In this instance there was one thread which I archived using User:Equazcion/OneClickArchiver because its only reason for not being archived was someone had added a note to a 6 year old section asking if it should be deleted. — Makyen (talk) 00:53, 31 March 2014 (UTC)

Care to take a look at Help talk:How to fix your signature? -- Kndimov (talk) 01:02, 31 March 2014 (UTC)

I have added a HTML comment to both of the drop in text sections above. I have also added the same text to what I will normally use (minus the carriage returns used above to prevent the <pre>...</pre> sections from being very wide). The comment is to explain to future editors that the time-to-archive should be changed in two locations. I used to have a much shorter comment within the MiszaBot config template. However, I found that having a comment within that template caused lcSB3 not to archive the page. After I reported the issue, the bot was "fixed" by having lcSB3 remove all comments from within its config template. While that solution is not perfect it is better than having the bot not archive, with no error reported, when there is a comment in its configuration template. I forgot to add back in a comment into my permanent clipboard set that explains the relationship between the time in {{User:MiszaBot/config}} and {{Archives}}. Sorry about that.

Note: The added comment starts on the same line as the end of the {{Archives}} template because the position of the table of contents (TOC) on the page is sensitive to extra lines between the templates. There are already a couple of things included in the drop-in text to make the TOC line up exactly with the archive box. Without previewing the page (in each instance) resulting from what is dropped in, I did not want to possibly introduce something that might cause the TOC to move down the page by some extra white-space. — Makyen (talk) 03:46, 31 March 2014 (UTC)

Oops, sorry, I missed the second age field. If I ever use this again I'll probably delete the comments when I set up the archiving bot. I think I'll just leave the title (<!--lcΣB3 2 years:-->) -- Kndimov (talk) 20:59, 31 March 2014 (UTC)