Talk:EMD GP38-2

There was never a railroad or corporation called Family Lines System. It was a marketing scheme of Seaboard Coast Line Industries for its Railroads. Those Family Lines GP38-2s in the roster will be moved to their proper railroads, a Family Lines paint will be placed as a note.

Working on the GP38-2 roster today, lots of errors. Checking A J Kristopan's EMD Serial Number page for corrections.

Saudi Arabia Government Railway had one GP38-2 #2000.

Total units will likely change. --SSW9389 (talk) 14:17, 20 July 2008 (UTC)

Using A J Kristopans Serial Number page found 1,799 GP38-2s for U S railroads and industrial operations. The production total on The Diesel Shop website is 1,801 units. Can't find the two unit discrepency. If anyone does, please advise. Found 156 Mexican GP38-2s for railroads and the one industrial operation. Found 257 Canadian GP38-2s for the railroads and one industrial operation. The Canadian figure includes the 51 wide nose GP38-2s for Canadian National. Found one GP38-2 exported to Saudi Arabia. My verified count for GP38-2s is 2,213 units. --SSW9389 (talk) 16:56, 20 July 2008 (UTC)

Who had the first GP38-2? I note that EMD built GP38-2s in January 1972 for L&N, MP, Southern, and Vermont. The first Canadian GP38-2 was for CN in November 1972. --SSW9389 (talk) 17:00, 20 July 2008 (UTC)

The report of the demise of the Texas Gulf Sulphur units appears to be premature. See http://www.trainweb.org/oldtimetrains/industrial/ont/kiddcreek.htm for some photos of the units in service.

Ongoing unexplained value changes without references by 211.250.168.108
It appears to be possible vandalism. No explanation. The user keeps re-instating his changes with no explanations or sources whatsoever despite having been reminded on their talk page and in edit summaries. Graywalls (talk) 21:39, 31 December 2019 (UTC)
 * Well, Foster (1996) gives a total production figure of 2188, which isn't either number. McDonnell (2015) gives 2222, which is the number the IP disagrees with. Using McDonnell and Wilson (2017), it may be possible to resolve discrepancies. Usually, the issue is export locomotives, Canadian-built units, or confusion over units sold to another railroad. Wilson doesn't give a production total, but he does enumerate locomotives for most Class Is. Mackensen (talk) 21:55, 31 December 2019 (UTC)

Looking at this further, the early December version of the roster claimed a total of 2222 but added up to 2211. Wilson (2017) agrees with the IP that Clinchfield had eight locomotives. For the Seaboard Coast Line Wilson says 64: 500–555, 6047–6050, 6053, 6054. Also, Wilson does give a production total: 2208. Mackensen (talk) 22:07, 31 December 2019 (UTC)

Robert Sarberenyi has the 2213 figure, including the GP38-2W. I expect most sources will commingle the two. Mackensen (talk) 22:14, 31 December 2019 (UTC)
 * I don't consider Robert Sarberenyi's railfan page a reliable source through the lens of WP:RS and this is more so when other published reliable source's data don't exactly agree with it. Graywalls (talk) 09:29, 1 January 2020 (UTC)

Rob Sarberenyi's GP38-2 data can be serial number verified using EMD documents. --SSW9389 (talk) 07:35, 3 January 2020 (UTC) But, if the data is only of interest to some rail fan and it has not been of enough interest to have been reliably published, the value of inclusion is questionable. We should evaluate if it is of interest to any interest or values to anyone but rail fans. If the answer is no, I opt for omission. I think the information is more appropriate to remain on rail fanning sites, not wikipedia. Graywalls (talk) 08:30, 3 January 2020 (UTC)
 * Just because something is published does not necessarily make it accurate. A serial number check is not everyone's cup of tea, but it does spell out accuracy of the data. --SSW9389 (talk) 12:11, 3 January 2020 (UTC)
 * Not just published, but RELIABLY published, per WP:RS and evaluation of reliability of sources per WP:COMPETENCE. We require verifiability, not truth, although, sometimes the truth is never available. If there's a discrepancy between reliably published source(s), and some dude on the internet, we go with the reliable sources. WP:VNT Graywalls (talk) 03:43, 4 January 2020 (UTC)
 * Well, it's been reliably published in several different books, but there's some disagreement. The variance isn't hugely significant--a dozen locomotives (save for Foster, and that work focused on spotting detail). Mackensen (talk) 12:16, 3 January 2020 (UTC)
 * What I'm saying is that, let's just omit this unimportant details altogether. 03:44, 4 January 2020 (UTC)

BNSF
This should go without saying, but BNSF was not an original owner. GP38-2 production ended in 1986, while BNSF was formed in 1995. I'm getting sick and tired of having to revert obvious factual inaccuracies repeatedly introduced by a certain IP editor. Trainsandotherthings (talk) 15:09, 12 November 2021 (UTC)
 * Roster lists may qualify as an attractive nuisance. I'm really not sure what's to be done about them. I'm coming around to the idea that they should be excluded altogether, with major or otherwise notable owners mentioned inline (consider EMD F40PH; there's still plenty of discussion of the important owners even if you removed the roster). F40PH may be a bad example inasmuch as we have good sources for the total build and they don't disagree. Mackensen (talk) 15:16, 12 November 2021 (UTC)
 * Honestly, I agree that they are a nuisance. They're a magnet for unsourced trivia and nonsense, and why should we bother hosting them when sites like the diesel shop already do? When I wrote GE 25-ton switcher, I intentionally excluded a roster list because as you can see it is ridiculously long (over 400 entries). In general, at minimum I think notes sections should be deprecated or seriously cut down, as they are basically cruft electromagnets and I question if they count as encyclopedic. It might be a good idea to launch another discussion at WP:TRAINS on the topic and see if we can get a general consensus to deprecate roster lists in general. I know they have their defenders, though. Trainsandotherthings (talk) 15:22, 12 November 2021 (UTC)

Unsourced Value Changes by 59.13.222.13
Hello, there have been some unsourced value changes by 59.13.222.13 and I believe that this is likely incorrect, but would like to come to a consensus as I do not have access to the source linked for this section before reverting the changes. This could be the same editor as mentioned by user graywalls, however it is hard to know. Input is appreciated. Jcody21 (talk) 20:48, 19 February 2024 (UTC)