This is an
essay. It contains the advice or opinions of one or more Wikipedia contributors. This page is not an encyclopedia article, nor is it one of
Wikipedia's policies or guidelines, as it has not been
thoroughly vetted by the community. Some essays represent widespread norms; others only represent minority viewpoints. |
By the end of 2013, the Page-Reformat Crisis had become a major problem, with some templates flooding the wp:Job_queue(s) with so many pages triggered, when updating the templates, that it has taken over 6 weeks(!) to update all related pages (which formerly ran within 2-3 days in 2008). The prior, slow Edit-preview Crisis, which used 12-45 seconds to preview changes to most major articles, was solved, in April 2013, by fast-cite templates, now using Lua script in wp:CS1 Module:Citation/CS1, plus fast Lua-based Module:Navbox and Module:Infobox or Template:Weatherbox (etc.) to cut edit-preview of pages as 3x-4x times faster than February 2013. Meanwhile, the epidemic growth of ever-more templates (plus Lua modules) has bogged down the page-reformat queues with millions of jobs to re-re-re-reformat the same pages for multiple changes to whichever megatemplates they are using. A recent victim has been Template:Convert, transitioned from 3,000 tiny subtemplates, to instead use a gargantuan Lua Module:Convert (at 02:15, 11 December 2013) but still reformatting the final 7,000 of the 554,000 related pages, almost 21 days later.
Potential solutions to page-reformat crisis: There are several possible options, but Step 1, is to beware people imagining, "It's not a problem" (no), because a delay of 2-to-7 weeks to show results of updated templates is too long, even if people try to rationalize the mega-slow results. You might recall, when the page " Canada" edit-preview ran 28 seconds or " Israel" ran 42 seconds, some people were saying not-a-problem and recommending to edit each major article by section only, but no longer seeing the whole page reformatted to proofread for typesetting. So, such solutions by lowered-expectations are a reluctant option, but would treat this website as a miserly operation. Instead, let's try to gain major performance improvements:
Fortunately, most templates rarely change major features, and so people have not been complaining (much), as they wait for a new option to appear in the final 7,000 pages within a half-million. However, as we upgrade to have smarter, auto-correcting templates, rather than issuing 1950s-style error messages ("ERROR: Invalid data; DOES NOT COMPUTE") as busy work to hand-edit, there is the need to re-install system-wide waves of auto-correcting updates (see: " wp:Autofixing cites"). In those cases, a high-priority of reformatting could post results weeks sooner, to check for unexpected consequences, and then re-release the next wave of smarter template within the week. Also, the flood caused by wp:data hoarding has multiplied thousands of minor pages with tedious data-table templates. Anyway, this is just a quick overview of the crisis, where corrections or updates to some major templates do not post current data for over 6 weeks of reformatting.
This is an
essay. It contains the advice or opinions of one or more Wikipedia contributors. This page is not an encyclopedia article, nor is it one of
Wikipedia's policies or guidelines, as it has not been
thoroughly vetted by the community. Some essays represent widespread norms; others only represent minority viewpoints. |
By the end of 2013, the Page-Reformat Crisis had become a major problem, with some templates flooding the wp:Job_queue(s) with so many pages triggered, when updating the templates, that it has taken over 6 weeks(!) to update all related pages (which formerly ran within 2-3 days in 2008). The prior, slow Edit-preview Crisis, which used 12-45 seconds to preview changes to most major articles, was solved, in April 2013, by fast-cite templates, now using Lua script in wp:CS1 Module:Citation/CS1, plus fast Lua-based Module:Navbox and Module:Infobox or Template:Weatherbox (etc.) to cut edit-preview of pages as 3x-4x times faster than February 2013. Meanwhile, the epidemic growth of ever-more templates (plus Lua modules) has bogged down the page-reformat queues with millions of jobs to re-re-re-reformat the same pages for multiple changes to whichever megatemplates they are using. A recent victim has been Template:Convert, transitioned from 3,000 tiny subtemplates, to instead use a gargantuan Lua Module:Convert (at 02:15, 11 December 2013) but still reformatting the final 7,000 of the 554,000 related pages, almost 21 days later.
Potential solutions to page-reformat crisis: There are several possible options, but Step 1, is to beware people imagining, "It's not a problem" (no), because a delay of 2-to-7 weeks to show results of updated templates is too long, even if people try to rationalize the mega-slow results. You might recall, when the page " Canada" edit-preview ran 28 seconds or " Israel" ran 42 seconds, some people were saying not-a-problem and recommending to edit each major article by section only, but no longer seeing the whole page reformatted to proofread for typesetting. So, such solutions by lowered-expectations are a reluctant option, but would treat this website as a miserly operation. Instead, let's try to gain major performance improvements:
Fortunately, most templates rarely change major features, and so people have not been complaining (much), as they wait for a new option to appear in the final 7,000 pages within a half-million. However, as we upgrade to have smarter, auto-correcting templates, rather than issuing 1950s-style error messages ("ERROR: Invalid data; DOES NOT COMPUTE") as busy work to hand-edit, there is the need to re-install system-wide waves of auto-correcting updates (see: " wp:Autofixing cites"). In those cases, a high-priority of reformatting could post results weeks sooner, to check for unexpected consequences, and then re-release the next wave of smarter template within the week. Also, the flood caused by wp:data hoarding has multiplied thousands of minor pages with tedious data-table templates. Anyway, this is just a quick overview of the crisis, where corrections or updates to some major templates do not post current data for over 6 weeks of reformatting.