It is customary to periodically archive old discussions on a talk page when that page becomes too large. Bulky talk pages may be hard to navigate, contain obsolete discussion, or become a burden for users with slow Internet connections or computers. Notices are placed at the beginning of the talk page to inform all editors of an archive.

Regular articles are not archived because previous versions may be seen in the page history; the practice of archiving is unique to talk pages and noticeboards, which often swell to great length.

Archiving one's own user talk page is optional; some users simply blank the page, as the history is kept available for future reference, but this is not considered the best practice (as it makes things more difficult to find and link). Removal of content from your user talk page, such as warnings posted by others, is considered evidence you have seen the content; this is true whether the removal was manual or automatic.

The talk page guidelines suggest archiving when the talk page exceeds 75 KB (or 75,000 bytes), or has multiple resolved or stale discussions. However, when to archive, and what may be the optimal length for a talk page, are subjective decisions that should be adapted to each case. For example, ongoing discussions and nearby sections they reference should generally be kept intact.

There may be circumstances where it would be useful to keep older discussions present on a talk page, to avoid the same issues being repeatedly raised. However, this situation can be better addressed by use of the ((FAQ)) template.

Technical overview

The preferred method for archiving a talk page is to copy the older sections to a series of subpages, and remove the sections from the original page. This can be performed either by hand or automatically by a bot.

There are two common methods of naming archive pages. The most common, and easiest, is sequentially numbered archives. Archive pages should be named as follows: take the name of the talk page, and add "/Archive #", where "#" is the number of the archive. Note that the word Archive has a capital A, there is a space before the number, and there are no leading zeros. Templates like ((Archive list)) rely on this naming convention to automatically find the archive pages.

For example:

The second common method of naming archive pages is by year and month. With this method archive pages are typically named Talk:YourPage/Archives/2013/June, Talk:YourPage/Archives/2013/July, etc. Some things are more complex with this method of naming. When using this naming scheme the subpages should use /Archives/ not /Archive/ as some of the navigation templates rely on Archives being plural.

Remember to use the correct namespace – the part before the colon (:) – when archiving your own user talk page. It should start with "User talk:", not "Talk:".

Are there any existing archived talk pages?

Most of the methods here rely on checking if any archived talk pages already exist, and placing the new sections to be archived either into the most recent archive, or in a new archive numerically sequenced one above the existing archive. There is a search to check if any subpages exist under a talk page (and archived pages are usually placed in subpages of the existing talk page). In the search bar type:

Special:PrefixIndex/Talk:Pagename

For example

Special:PrefixIndex/Help talk:Archiving a talk page will display the talk page to this help page and all its archives.

Manual archiving

Cut and paste procedure

  1. Edit the talk page, copying material you wish to archive to your computer's clipboard.
  2. Create a new archive page at Talk:Example page/Archive N, where N is the lowest number for which no archive exists. (You can create the page by searching for the desired title and clicking the red link in the search results, or by editing the URL bar to point to the desired location then clicking the "Create" button.)
    • Alternatively, you may occasionally wish to create a topical archive, in which case the archive should be named according to the topic. For example Talk:Example page/Place of birth debate.
  3. Paste the clipboard contents into this archive and add ((Talk archive)) (or ((Automatic archive navigator)) to enable navigation between archives) to the top of the code. Publish the new archive.
  4. Delete copied material from main talk page with edit summary mentioning the name of the new archive.
  5. If an archive box doesn't already exist on the main talk page, add the line ((Archives)) below the WikiProject tags and publish. (This will not be necessary if the main talk page already contains the line ((Talk header)), since this template automatically contains a list of archives if they are named in the standard way.)
  6. If there is already an archive box on the relevant talk page, you may want to add to it the link to the new archive.

For a more detailed set of instructions, see /Manual archiving.

Semi-automated cut and paste procedure

OneClickArchiver and Archy McArchface are userscripts designed to simplify the above manual procedure. Instead of manually copying and pasting the text, these scripts allow you to simply check which sections you wish to archive and specify a page to archive the selected sections to. Factotum is not a dedicated archiver script but can move a section to another specified page, which can be a talk page archive.

Other manual procedures

Other manual procedures were once considered equal alternatives with the cut and paste procedure described above. Over time both methods fell out of use and are generally not used. If you have used the page move procedure those pages do not qualify for speedy deletion by user request.

Automated archiving

For plain and simple automated archiving, see Help:Archiving (plain and simple).
Shortcut

ClueBot III and lowercase sigmabot III can automatically create archives for any discussion page (one that has "talk" in the namespace), by moving sections to a subpage when the section has received no comments for a specified period of time. Both bots can archive sections into simple sequentially numbered archives (e.g. YourPage/Archive 1, YourPage/Archive 2, ...), or archive according to date (e.g. YourPage/Archives/2013/June, YourPage/Archives/2013/July, ...).

See the ClueBot III page and the lowercase sigmabot III page for detailed instructions on setting up these bots. Those pages explain each of the parameters used, such as archiveheader, minthreadsleft, etc. The examples below are also available for use. All of the examples here are intended for you to be able to copy them while viewing this page and paste them into the edit box on your target page without requiring any modification to have a working configuration.

Choosing a bot

ClueBot III and lowercase sigmabot III, the two bots in current use for automated archiving, are similar in their behaviour, but have some differences in terms of features they implement and limitations on their use:

lowercase sigmabot III ClueBot III
Number of talk pages in use on ~37,000 ~10,600
Index of archives? requires additional bot Yes
Automatically repair links to discussions when archiving? No Yes (example)
Initial delay before first archive Less than 24 hours (typically begins runs at 00:00 UTC daily) Up to several days (example)
Limitations Will not archive sections without signatures Not suitable for pages linked to by a number of pages in the order of thousands (for performance reasons)

The following are example configurations to set up automatic archiving of an article talk page or user talk page. These configuration templates do not produce any visible output, so you should generally also include a template such as ((Archives)) to provide readers with links to search and navigate the archives. These templates may also be configured with the parameters bot and age to include a message notifying readers that automatic archiving is taking place. For example: ((Archives|age=90|bot=ClueBot III)).

Sequentially numbered archives

Sequentially numbered archives
Lowercase sigmabot III ClueBot III 
((User:MiszaBot/config
| algo=old(90d)
| archive=((SUBST:#titleparts:((SUBST:FULLPAGENAME))))/Archive %(counter)d
| counter=1
| maxarchivesize=75K
| archiveheader=((Automatic archive navigator))
| minthreadsleft=5
| minthreadstoarchive=2
))

((User:ClueBot III/ArchiveThis
| age=2160
| archiveprefix=((SUBST:#titleparts:((SUBST:FULLPAGENAME))))/Archive
| numberstart=1
| maxarchsize=75000
| header=((Automatic archive navigator))
| minkeepthreads=5
| minarchthreads=2
| format= %%i
))

Both of the above examples will create archive pages, as needed, starting with TalkPage/Archive 1 and continuing with TalkPage/Archive 2, etc., when the current archive page grows to more than 75,000 bytes. Only sections with dates that are older than 90 days (2,160 hours) will be archived, although the five sections with the most recent contributions will be retained on the talk page no matter how long ago the last contribution was added. If archives for the page already exist, then set the counter/numberstart parameter to the next archive number in the series.

Warning: the magic word ((SUBST:FULLPAGENAME)) is replaced by the talk page name when the template is saved to a talk page. However some punctuation characters that can appear in a page name ("&') are replaced by HTML character codes that are not recognised by the bots (see mediawikiwiki:Manual:PAGENAMEE encoding#PAGENAME for details about these and other characters). For example ' is replaced with ' (so if the bot is to work then ' must be replaced with '). This can be circumvented by using ((SUBST:#titleparts:((SUBST:FULLPAGENAME)))) to automatically sort the encoding.

Archives by year and month

Archives by year/month
lowercase sigmabot III ClueBot III

Note: In this configuration lowercase sigmabot III puts threads into archives based on the last date in each thread. Here is an example of it splitting an archiving task of 185 threads into 54 YYYY/Month based archives.

Note:Each time ClueBot III runs it places threads in a single archive which is the YYYY/Month of (NOW - age). If you are starting archiving with a backlog this will result in a single file with all of your old threads older than age. ClueBot III does not appear to run every single day. Thus, even when archiving has been running for a while, some threads may end up in an archive that is one month later than when they were last edited.

((User:MiszaBot/config
| algo=old(90d)
| archive=((SUBST:#titleparts:((SUBST:FULLPAGENAME))))/Archives/%(year)d/%(monthname)s
| minthreadsleft=4
| minthreadstoarchive=2
))

((User:ClueBot III/ArchiveThis
| age=2160
| archiveprefix=((SUBST:#titleparts:((SUBST:FULLPAGENAME))))/Archives/
| minkeepthreads=4
| minarchthreads=2
| format=Y/F
))
Both of the above examples will create archive pages, as needed, with names similar to YourPage/Archives/June, YourPage/Archives/July, etc. Only sections with dates that are older than 90 days (2,160 hours) will be archived in these examples.

Bot and config details

The configuration template for lowercase sigmabot III is still named ((User:MiszaBot/config)) even though the original bots, MiszaBot I, MiszaBot II, and MiszaBot III, are no longer active.

The best age for threads to be archived will vary, depending on talk page activity. The purpose of archiving article talk pages should be to remove stale discussions, not to stifle discussion. Article talk page threads should not typically be archived in less than 30 days except for very busy talk pages.

The maxarchivesize/maxarchsize parameter determines how large each archive is allowed to grow before a new archive is started. Keeping this value larger than a typical talk page allows the archives to be browsed more easily, but increases page load times. Each individual archive should not be larger than 512kB, because this may cause accessibility problems for some devices. Because a large batch of threads can "overshoot" the maxarch[ive]size parameter, the parameter should always be set lower than the maximum acceptable archive size.

Category:Pages where archive parameter is not a subpage

One limitation of the way automatic archival instructions are set up is how page moves will easily break the functionality. For several reasons, the bots will generally only archive to subpages; that is, the archive must reside as one or more subpages to the talk page in question.

When a page is renamed (moved) and the user forgets to also update the bot instructions, this means that no further archiving will take place (until the bot instructions are updated manually). To facilitate identifying cases where archiving is interrupted, the Category:Pages where archive parameter is not a subpage was created.

If your talk page isn't archiving properly, you might want to check this category. If your page is included, this indicates that the specified archive pages are incorrect.

Templates

For archive pages

A template should be placed at the top of each archive page indicating that it is an archive. The simplest such template is ((Talk archive)). If the page is using sequentially numbered archives, you can instead use ((Automatic archive navigator)) which will add navigation links to earlier and later archives (example).

ClueBot III and Lowercase sigmabot III will automatically place ((Talk archive)) on archive pages they create. They can be configured to use other templates using the header and archiveheader parameters, respectively. For example (for ClueBot III), |header=((Automatic archive navigator)).

For talk pages

There are many templates that can be placed on talk pages to make the page's archives visible and accessible to editors. These can be broadly divided into "boxes" (which generally appear floating to the right of the table of contents), and banners. Some pages use one or the other, and some use both. Templates dedicated to archiving should generally be placed last, after all other talk templates, but before the discussions proper.

Archive link boxes

On regularly archived talk pages, it is useful to have an "archive box" template. Common usage is to place the archive box below other header templates and before the first section heading so that the box appears to the right of the table of contents. If the page already has a manual archive box, add the link to the new archive page to it. If there is no archive box yet, you may want to set one up.

The most common archive box template is ((Archives)). It has a significant number of options. The auto option causes the box to automatically find your archive page list. If auto=yes or auto=short the links are shown as "1, 2, 3". If auto=long the links are shown as "Archive 1, Archive 2, Archive 3"). If the archives do not follow the standard naming scheme (/Archive #), or date labels are desired a manual list can be inputted in the first unnamed parameter as seen in example 2 below. There are also options which show a search field, permit the box to collapse/expand, start the box collapsed/expanded, etc. Many of these are shown in example 4 and all of them are explained in more detail at Template:Archive/doc.

Example 1

For a box with an automatically generated archive list with a search bar. ((Archives)) can be used without any extra parameters.

Example 2

For a box with a manual archive list and a search bar ((Archives|List of archives)) can be used. For the box to the right the code is: 


((tlp|Archives|
*[[Help_talk:Archiving_a_talk_page/Archive_1|2003–2006]]
*[[Help_talk:Archiving_a_talk_page/Archive_2|2006–2007]]
*[[Help_talk:Archiving_a_talk_page/Archive_3|2007–2009]]
*[[Help_talk:Archiving_a_talk_page/Archive_4|2008–2014]]
*[[Help_talk:Archiving_a_talk_page/Archive_5|2014–2019]]))
Example 3

For a box with both an automatic and manual list ((Archives|List of archives)) can be used. The example below for Template talk:Archives where the archives for the old ((archive box)) template are shown use the following code: ((Archives|auto=long|'''Archive box archives''':<br>[[Template talk:Archive box/Archive 1|Archive 1]], [[Template talk:Archive box/Archive 1|Archive 1]]))

Example 4

This template just demonstrate a lot of the parameters. Most of them are quite self explanatory. 

((archives
| auto        = short
| search      = no
| root        = Help talk:Archiving a talk page
| collapsible = yes
| collapsed   = yes
| style = background-color:#FFAAAA; width:140px
| title       = These are the archives!
| image       = File:Crystal Clear app file-manager.png
| image-size  = 60px
| bot         = Lowercase sigmabot III
| age         = 1
| units       = year
))

Automatic archive boxes for year/month archives

Automatic archive boxes for Year/Month archives are a bit more complicated.

The most common method is by using the ((Archives by months)) template to generate links to your existing archives. It produces a live link for each existing archive month within a year. It is commonly used in something similar to: 

((Archive box |search=yes |collapsible=yes |
bot=ClueBot III |age=90 |
title=[[Help:Archiving a talk page|Archives]]<!--
--> ([[((#titleparts:((TALKPAGENAME))|1))/Archive index|index]])|
((Nowrap|'''2004''': ((Archives by months|2004))))
((Nowrap|'''2005''': ((Archives by months|2005))))
((Nowrap|'''2006''': ((Archives by months|2006))))
((Nowrap|'''2007''': ((Archives by months|2007))))
((Nowrap|'''2008''': ((Archives by months|2008))))
((Nowrap|'''2009''': ((Archives by months|2009))))
((Nowrap|'''2010''': ((Archives by months|2010))))
((Nowrap|'''2011''': ((Archives by months|2011))))
((Nowrap|'''2012''': ((Archives by months|2012))))
((Nowrap|'''2013''': ((Archives by months|2013))))
((Nowrap|'''2014''': ((Archives by months|2014))))
((Nowrap|'''2015''': ((Archives by months|2015))))
'''2016''': ((Archives by months|2016))
))

The above code produces the archive box to the above right. The example here has no month names which are links because there are no archive files for this page. You can, of course, delete the lines for the years you don't want.

The default for the ((Archives by months)) template is that archive page names are in the format YourPage/Archives/2013/June, YourPage/Archives/2013/July, etc. However, ((Archives by months)) allows you to specify various naming conventions for your archives. Archives can have an arbitrary prefix by using the archprefix parameter (e.g. archprefix=Archive/ instead of Archives/). Months can be in four formats: A) the default uses January, February, March, etc.; B) abbrev3=yes months are Jan, Feb, Mar, etc.; C) numbers=yes months are 1, 2, 3, ... 12; D) numbers0=yes months are 01, 02, 03, ... 12.

If you are using ClueBot III, it has a couple of additional options for creating an archive box. These involve the use of either the archivebox or the index parameters. See the ClueBot III page for more information.

Archive link banner

Instead of putting the archive links in a box that floats to the right or left, some prefer the links in a banner like ((Talk header)). Yearly and alphabetic archives are automatically detected in addition to regular sequential archives. A search box is by default included and if the page is archived by bots an auto archiving notice can be included by specifying |archive_age= in days.

((Talk header))

The archive links show up in a row at the bottom of the above banner. For a real world example see Talk:Proxima Centauri.

Archive indexing

Having an archive index makes finding old discussions on a given topic easier, particularly in pages with many archives, or when the archives are of considerable size. An archive index, when one exists, is typically stored on the page Talk:YourPage/Archive index. While it is possible to generate an index by hand, the /Archive index page is normally generated automatically. There are two bots which automatically create indexes: Legobot and ClueBot III.

Legobot, formerly HBC Archive Indexerbot, can create an archive index page based on a set of archives. See the bot's instructions for details on how to set up archive indexing. Legobot generated archive indexes are the most commonly available type of index. To set up Legobot archiving you have to create your Talk:YourPage/Archive index page and need to place:
<!-- Legobot can blank this -->
on top of the index page. The following is a Legobot configuration that works with numbered archives which you can place on the page you are archiving:

Example Legobot configurations
Numbered archives Archives/Year/Month 
((User:HBC Archive Indexerbot/OptIn
|target=/Archive index
|mask=/Archive <#>
|leading_zeros=0
|indexhere=yes
))

((User:HBC Archive Indexerbot/OptIn
|target=/Archive index
|mask=/Archives/<year>/<month>
|indexhere=yes
))

ClueBot III automatically creates an archive index when archiving. ClueBot III stores the archive indexes it creates on unique pages in its user space. The archive index for the page you are archiving can be transcluded onto your Talk:YourPage/Archive index page by putting the following text in that page:
((User:ClueBot III/Master Detailed Indices/((NAMESPACE)):((BASEPAGENAME))))

Archive searching

If there are several archives, it is easy to search them all at once using the optional search parameter prefix. For example, suppose we want to begin a discussion about adding a security section to the Linux article. Before bringing up the topic we can do the following search: security prefix:Talk:Linux/. This will look in all the subpages of the Linux article in its Talk namespace.

The templates ((Archives)), ((Archive box)), and ((Talk header)) have a parameter search=yes which provides a helpful search box on the talk page to automate searching, as do independent search templates such as ((search archives)). It is recommended that when a page is archived, such a template be put on the surviving talk page to simplify access to past discussions.

Continuing discussions

If a thread has been archived prematurely, such as when it is still relevant to current work or was not concluded, unarchive it by copying it back to the talk page from the archive, and deleting it from the archive. Do not unarchive a thread that was effectively closed; instead, start a new discussion and link to the archived prior discussion.

Example pages

Here are actual discussion pages on Wikipedia where you can study the code to see how archive pages were created.

See also

Categories