R&S Example page/Verification cycle

Verification cycle

In a discussion on Harry's talk page, Ahasuerus wrote:

My concern is that the verification cycle usually runs something like this:

1. Skeleton data was originally entered from unattributed sources, sometimes years ago and often by robots, and may include incorrect or misleading information
2. Secondary sources are subsequently used to verify the record, e.g. "Page count from Amazon.com and may be unreliable", "Binding unknown, but either hc or tp based on OCLC", "Cover art attributed by the Locus Index" or "First edition according to Tuck"
3. Finally, primary verification occurs, at which point we delete references to secondary sources re: "objectively verifiable" fields like binding and the page count -- there is no reason to state what OCLC or the Locus Index says about the page count once the book has been physically verified.

On the other hand, some information originally entered from secondary sources is retained even after physical verification, typically if the publication is missing some information or if there is reason to believe that some of the stated information is invalid. For example, the Notes field may say "Stated second printing, but actually sixth printing according to the author's Web site" or "Publication date not stated, but 1971 according to Reginald-1".

Based on this cycle, a reference to a secondary source in a verified publication generally means that there is missing or incorrect information in the book, so if we are going to leave references to secondary sources, it's important to clarify the nature of the secondary information. ... Ahasuerus 15:26, 16 August 2009 (UTC)

I copy this here because I do not strictly follow the above described pattern, and wonder if I should.

Specifically, if I primary verify a publication, I do not usually remove secondary sources mentioned in notes. If the data from those sources contradicts what I find in the pub, I modify the note to make this clear. If it confirms the physical pub, I usually make no change to the existing note. If it confirms data not in the pub (such as a date or artist credit) I usually add a note that no such date or credit is in the pub itself. But I pretty much never simply remove such a note.

Moreover, I do secondary verifications, particularly OCLC and Locus verifications, on pubs already primary verified, and tend to add notes like "Price confirmed by Locus1" or "Page count differs from that given by OCLC".

When creating a pub record from secondary sources, I now tend to add notes indicating the source of each data element, particularly if I assemble the data from multiple sources (OCLC+ABE+Amazon+Locus, say). Do people generally think that secondary source notes ought usually to be erased when a pub is primary verified, and that the retention or later addition of them should signal a problem, rather than a confirmation? (I know Bill Longley prefers not to note sources at all in many cases, but I think that is a minority opinion.) -DES ^Talk 16:16, 16 August 2009 (UTC)

I certainly don't add notes confirming that a secondary source is right, I would if it was wrong. As for my not adding notes as to my sources - well, if I'm concocting a frankenpub that we didn't have before from secondary sources I'd rather have it treated as a step 1 skeleton pub than imply that any of the sources were actually more reliable than others. Or to have to do that much typing - I might have used 7 or 8 sources at times and still not finished finding the contents of an anthology, for instance. But at least then I'll indicate it has "partial contents from reviews", but I'm not going to list every author site that said "My story X is included in Anthology Y". BLongley 19:42, 16 August 2009 (UTC)

I'll try to list the classes of cases that we have run into with secondary sources:

1. The secondary source is incorrect. I believe the fact is worth recording in the Notes fields for major bibliographies, e.g. Tuck, the Locus Index or Reginald. On the other hand, I wouldn't record the fact that Amazon is wrong about the page count since it's pretty much their default mode of operations. However, more unusual situations may be worth noting, e.g. if they have the wrong cover scan displayed or if they list vaporware items.
2. The secondary source is less precise than what we see during the primary verification. For example, OCLC may list the book's dimensions, e.g. "22", which we may translate as "hc or tp, but definitely not a pb", but then when we verify the pub we know for sure whether it's a hc or a tp, so I see no value in retaining the OCLC data, which will only obfuscate the picture.
3. The secondary source matches the verification copy. For "objective" "what you see is what you get" fields like binding, price, ISBN and the page count, I see no value in stating that OCLC, Locus and Reginald all agree that the book is, indeed, a 250pp hardcover and that the stated price is $18.95. For data elements that publishers often get wrong and that we usually capture in Notes, e.g. the edition statement, it may be worth noting that, say, Reginald or Locus agrees that the book is indeed a first edition.
4. The secondary source provides information otherwise not available in the verification copy, e.g. it identifies the cover artist or the printing number. Naturally, we will want to keep this information and make a note of where it came from.

Having said that, these are just the guidelines that I use and I am sure other editors use somewhat different guidelines. There is no harm in retaining somewhat more (or less) secondary information in Notes -- as long as it's clear where the data came from -- and I doubt it would be practicable to come up with a universally acceptable standard. Ahasuerus 02:36, 17 August 2009 (UTC)