Walt at Random

I’m now just over halfway through the initial journal scan for GOA6 (8,000 of 15,676), so it’s a good time for a quick progress report.

Note that, as before, I sort journals by publisher before checking–because many multijournal publishers use the same templates for all journals, making it easier for me to find fee data and do article counts.

For GOA6, that means I’ve now checked through publisher PPCT and title Kimün. Revista Interdisciplinaria de Formación Docente. So far, the 2020 article count is 704,825, and that will almost certainly go up slightly. The 2019 total for this set of journals is 586,583 articles. [Yes, the substantial gain for 2020 appears to be legit: for one thing, quite a few MDPI journals saw substantial increases in articles in 2020. And there are quite a few more journals.

Last year, that range of publishers included 7,155 journals, which published 569,645 articles in 2019. So there’s a net gain of 625 added journals so far.

Note that the second remainder of last year’s journals accounted for 281,668 2019 articles, not quite half as many as the first half. So we could wind up with a million articles, but certainly not 1.4 million or close to it. (I’d say a million is probable unless rechecking shows major problems in the figures.)

For this group of 3,200 journals–ignoring the first 4,800–problematic journals include 129 malware cases (that’s out of 3,200–the malware numbers are bad this year, but the latest 1,600 didn’t make them much worse, adding 43), 146 unreachable/unworkable, as well as 47 that had to be reached at new addresses (entirely because DergiPark moved from .gov to .org). These will all be rechecked.

Looking more closely at the malware cases for these 3,200 journals [thus including the fourth report], there are nine security-certificate problem, one spyware, seven malware in general, 18 phishing–and 94 Trojans. Looking at countries in this 20%, I see 70 in Indonesia, 21 Ukraine, 9 Serbia, 3 each in Brazil and Poland, two each in Colombia, Kenya and Turkey, and seventeen singletons.

Now, on to the next 1,600… and I might stop providing problematic-journal details, if the reports continue at all. Perhaps the most striking thing in this segment is that the 58 largest journals in the latest 1,600–mostly from MDPI, but with a few from Nature Publishing Group and others–went from 129,856 articles in 2019 to 184,357 in 2020.

I’m now just over 40% of the way through the initial journal scan for GOA6 (6,400 of 15,676), so it’s a good time for a quick progress report.

Note that, as before, I sort journals by publisher before checking–because many multijournal publishers use the same templates for all journals, making it easier for me to find fee data and do article counts.

For GOA6, that means I’ve now checked through Magazine House of Cancer Research on Prevention and Treatment; so far, the 2020 article count is 443,866, but that will almost certainly go up slightly. The 2019 total for this set of journals is 386,364 articles.

Last year, that range of publishers included 5,775 journals, which published 376,629 articles in 2019. So there’s a net gain of 625 added journals so far.

For this group of 1,600 journals–ignoring the first 4,800–problematic journals include 86 malware cases (that’s out of 1,600–the malware numbers are VERY bad this year), 88 unreachable/unworkable, as well as 23 that had to be reached at new addresses (entirely because DergiPark moved from .gov to .org). These will all be rechecked.

Looking more closely at the surprisingly high number of malware cases for these 1,600 journals, there’s one security-certificate problem, one spyware, six malware in general, 15 phishing–and 63 Trojans. Looking at countries in this 10%, I see one case each in Brazil and India, 54 in Indonesia (mostly Trojan, some phishing, almost all at academic institutions), one each in New Zealand, Pakistan and Romania, two in Russia, seven in Serbia, one each South Korea, Taiwan, Thailand and Turkey, one in the UK–and 14 in Ukraine, 13 of the from LLC “CPC “Business Perspectives”–that is, Trojan in the base software for all 13 of that publisher’s journals.

Looking back at the 174 xm cases in the first 4,800 journals, again by country, I see one each in Argentina and Bangladesh, four in Belarus, 43 in Brazil–35 of them from the same publisher, Conselho Nacional de Pesquisa e Pós-graduação em Direito (CONPEDI) (another root-software Trojan ); one from Chile, three from Croatia, one from Cuba, four from Ecuador, two from Germany, and 75 from Indonesia (that’s in addition to the 54 in the latest 1,600); then one from Iraq, two from Mexico, one from Moldova, six security-certificate problems from BRILL in Netherlands, one each from Pakistan, Poland and Portugal, ten from Romania, seven from Serbia, one each from Slovakia and South Africa, three from Spain, and one each from Ukraine and the US. (Portugal? That was a top-level domain issue, and I’m now ignoring these–most are .info and shouldn’t be flagged at all.

OK, way too much detail on malware issues, but they seem to be getting worse. Two years ago, before the pandemic, DOAJ and its contacts were able to correct nearly all malware cases. Not so last year, and I don’t know what will happen this year. At the end of this project for this year, I will send DOAJ a list of all journals with malware in both years and the suggestion that they be removed from the directory, possibly after one final attempt to get them to fix the problem

Now, on to the next 1,600 and the halfway mark. (Yes, I’m still on pace to be finished with the first pass in late April; I hope that continues.)