- FastCode (talk|edit|history|logs|links|watch) (XfD|restore)
We were in discussion of the close with Spartaz when he went on vacation. So we are apparently going to skip that step. I will say that the difficulty of getting a page listed has completely surprised me and raised my respect for WP. That said, I thought the delete discussion was going well and pretty much everyone had reversed their delete vote when it was closed. I think this deletion should be reviewed. Blwhite (talk) 16:25, 31 January 2011 (UTC)[reply]
- Sure, we'll happily review it. The deletion review is a great deal more likely to go in your favour if you will kindly provide a list of the sources you intend to use for this topic, including at least two that are not blogs or any other form of user-submitted content, are fully independent of FastCode and any associated people or corporations, and provide non-trivial coverage of the subject.—S Marshall T/C 17:32, 31 January 2011 (UTC)[reply]
- The appeal to the closing admin is here. Many of the sources in the article were either self published, not third-party, or only passing mention. The only two that I think really worked toward establishing notability per wp:GNG are below. Unfortunately they are not available online. When I saw them, I withdrew my delete !vote because I felt like this might be enough to satisfy gng. Spartaz apparently missed it before closing, and after getting a chance to read them thought that it still wasn't enough; you can see this on the discussion I linked to above. Here are the sources:
- Long, Brian & Swart, Bob, "Borland Developer Studio 2006 Reviewed", The Delphi Magazine, Issue 124, December 2005
- This devotes a few paragraphs to explaining what fastcode is and how some of its projects were included in the new version of this product. ErikHaugen (talk | contribs) 18:08, 31 January 2011 (UTC)[reply]
- A new release product overview, is naturally going to be summary in nature. But note that even though Delphi 2005 is generally considered the buggiest version since Delphi 4, the FastCode additions got more attention than the more than 1000 bug fixes. Blwhite (talk) 01:34, 1 February 2011 (UTC)[reply]
- Gabrijelcic, Primoz, "To Manage Memory", The Delphi Magazine, Issue 126, February 2006
- There is a half a page talking about what fastcode is and how it works in order to introduce a memory manager that came out of the fastcode project. ErikHaugen (talk | contribs) 18:08, 31 January 2011 (UTC)[reply]
- To be fair, the whole 8 page article is about FastMM and why it is better than the existing allocator from Borland, including how it grew out of the FastCode challenges. For those who are not familiar, FastMM is now the poster child of the FastCode project. But it was not the origin, or initial purpose of FC. FMM came as a natural result. Blwhite (talk) 01:34, 1 February 2011 (UTC)[reply]
- Endorse deletion The GNG wants significant sources in addition to third party sources. Neither "a few paragraphs" nor "a half page" are significant. No other coverage in five years? I'm guessing they also don't show anything about how Fastcode is extraordinary not WP:MILL. SchmuckyTheCat (talk)
- The rest of the discussion is where is the boundary for being an independent source. All the rest of the sources are familiar with the situation, and are therefore being counted as not independent.
- To clarify, it is as if an open source group started a project and wrote an installer intended to replace the MSI installer from MS. This new installer is compatible with existing MSIs and can be used directly. Having blogs and other sources from Borland advocating and recommending FastCode routines is the equivalent of MS engineers coming out and saying, "this new installer is better than anything that we can come up with. We recommend using it." I can't imagine that anyone would say that the MS sources weren't independent, as no one at MS was involved in the project. Yes, they benefit from it. And yes, they are biased, but the other way! It would be only natural to resist admitting that these open source guys can do better. In my mind, that makes it all the more impressive when they come out in favor of it.
- So it is in this case. There had been many attempts over the years to suggest improvements to Delphi and even to submit code changes to Delphi itself by the Delphi community. By and large, these were met with stone cold silence. FC itself, at first, received the same reception from Borland. So we see the inclusion of FastCode routines in Delphi as a major win for the project members, and the community as a whole. Conversely, we are confused by the finding that references to FastCode and FastMM from inside Borland are not independent.
- So yes, we have several references. But I want to get this question of whether Borland references count as independent resolved first so that we don't flood you with a bunch of useless information. For example, Steve Trefethen worked on incorporating Fastcode stuff into Delphi 2007. That is, he did not work on the FastCode project itself. In my mind, he is like the MS installer engineer excited about replacing the MSI installer.
- HTH. Blwhite (talk) 01:34, 1 February 2011 (UTC)[reply]
- I feel like the Trefethen source is a little too self-published to do much wrt GNG, even if it is deemed fully independent. ErikHaugen (talk | contribs) 07:28, 1 February 2011 (UTC)[reply]
- Overturn Sources were added after the Afd started, and from then on, all comments were in favor of keeping. This reflects the more final consensus. Shaliya waya (talk) 04:30, 1 February 2011 (UTC)[reply]
- Overturn In addition to the references listed above, FastCode is also explained in Marco Cantu's book, Mastering Delphi 2005, p. 157. The closing admin's comments, both on the AfD page and on his personal user page, consistently misrepresented the discussion. E.g., on AfD he stated that providing sources "doesn't appear to have happened" and on his user page he said that sourcing "wasn't discussed." Clearly, neither one was true. Although I don't think that AfD is the best way to request sourcing for an article, it did have the effect, in this case, of getting verifiable citations, and more were being added throughout the process. Because the close appears to have been based on factually incorrect assumptions and because WP:GNG was being actively addressed, I think this should be reconsidered. --Craig Stuntz (talk) 13:46, 1 February 2011 (UTC)[reply]
- Overturn Subject appears notable, citations appear legitimate, what's the problem?Oceansummer87 (talk) 19:22, 1 February 2011 (UTC)[reply]
- Overturn The GNG is careful to avoid saying what is meant by "substantial". I consider it to mean anything more than routine identification, rather than a fixed number of words or sentences. Certainly many articles and reviews about computer software & the like, even of clearly notable products, tend to be half a page or so. Merely the fact that something is chosen for such write up is notability, because most are not. Borland refs are not totally independent, but there's no reason not to think they are objective--that a major manufacturer includes something in its product is relevant, and its description is good source. Spartaz usually makes good decisions, but for a few of the most recent ones, he seems to have ignored improvements in sourcing during the discussion. DGG ( talk ) 20:57, 1 February 2011 (UTC)[reply]
- Overturn. I agree with Craig Stuntz's identification of the problems with the close. The reliability of the sources ought to be assessed, though, going forward. I'm not convinced the article is proved yet. --Bsherr (talk) 01:30, 2 February 2011 (UTC)[reply]
|