Walt at Random

Here’s how things are going for GOAJ2: Gold Open Access Journals 2011-2016 and The Countries of OAWorld 2011-2016 and related stuff (all linked to from the project home page), through February 28, 2018 [noting that most of the last day of each month is missing because of how statistics are done]:

• The dataset: 379 views, 72 downloads.
• GOAJ2: 1,452 PDF ebooks (and two paperbacks), plus 792 copies of chapters 1-7 (C&I 17.4)
• Countries: 517 PDF ebooks (no paperbacks)
• Subject supplement (C&I 17.5): 1,299 copies

Looking at the rest of gold OA (beyond DOAJ), Gray OA 2012-2016 has now been downloaded 2,261 times, while Gray OA 2014-2017: A Partial Followup, released in October 2017, has 838 downloads.

I’ve reached the 2/3 mark (more or less) in the Gold OA Journals first scan–7,200 down, around 3,500 to go–and thought I’d look at the rest of malware (so far), since I’ve encountered a few clusters beyond the many in Indonesia and those already reported in Romania.

There are 43 (so far)–and, notably, it appears that none of them was infected last year.

If you know of someone who might be able to check and fix some of these, it would be good for anybody who wants to keep their computer healthy.

The country counts: Bangladesh 1, Brazil 8, Bulgaria 1, Georgia 1, Iran 1, Italy 1, Malaysia 5, Mongolia 1, Morocco 2, Nepal 2, Peru 2, Poland 1, Portugal 1, Tunisia 1, Ukraine 7 (six in the same domain), UK 1, US 1.

A hard-to-read table is appended. If you want to work on these and find this useless, send me email (waltcrawford@gmail.com) and I’ll send back an .xslx file. When I finish the first pass, I’ll post a spreadsheet with all of the malware (at that point); I’m sure I’ll encounter a few more, although I hope “a few” turns out to be the right term.

In a post about the malware problem in far too many of Indonesia’s open access journals, I included a “PS” noting that Romania had the second largest number of infected journals in the first 4,000, although that number wasn’t very large (9)–and that Romania had the largest number of infected journals last year (11). I said I’d check the remainder of the Romanian gold OA journals and post a similar message with spreadsheets.

I’ve completed that scan (and I’m now not quite halfway through the first scan), and the story is that there are, in all, eleven Romanian OA journals flagged as malware by Malwarebytes Premium. That’s more than any other country except Indonesia, but it’s not a big percentage, and not really worth a spreadsheet. I provide the titles, URLs and publishers at the end of this post, in what is probably a clumsy table. (I believe that these 11 are entirely different than last year’s 11, but it’s possible that some secondary columns didn’t line up right, so I wouldn’t swear to that.[confirmed on 2/13: these are entirely different journals])

Other countries? A total of 14 of the first 5,340 journals–four in Brazil and one each in ten other countries.

Here’s the table, such as it is (if you want to work on these and find the table unusable, send me email at waltcrawford@gmail.com and I’ll send back an Excel spreadsheet):

If you know of someone involved with gold OA within Indonesia (probably working with universities), you might want to let them know about the situation described here. I normally just observe, measure and comment, but this case seems special enough to at least support action.

The Good News

Whether due to DOAJ‘s efforts or otherwise, the number of Indonesian gold OA journals within DOAJ grew rapidly during 2007, with 665 new journals (and a few apparently returned to the directory). I believe Indonesia now has more journals within DOAJ than any other country: 1,154 as of 1/1/2018.

The Bad News

Indonesia seems to have a problem with malware. Malwarebytes Premium, which I always run (alongside Windows Defender) after problems encountered with a few infected OA journals three years ago, finds an astonishing 286 Indonesia journals afflicted with malware–either the journal site itself or a site called by the journal site.

[How astonishing is that number? In the 2016 study, there were a total of 67 infected journal sites; so far, with almost half the journals checked, there are 21 outside of Indonesia.]

I do not ignore Malwarebytes warnings: I close the browser tab immediately. That means I don’t check the journal, and it won’t be included in the 2012-2017 Gold Open Access Journals report this summer–and, frankly, that any cautious user will avoid it.

I do recheck every malware-infected site during my second pass in April or early May (where I also recheck all other problematic sites–those unreachable, suspended, etc.–and the sites where it seems at least possible that late-2017 articles are being posted late). For journals that continue to be infected or problematic, I glean as much information as I can from DOAJ, but while I include these numbers in one or two “big numbers” tables, they’re simply excluded from the bulk of the report.

How to Help

People in contact with the publishers–mostly universities–could suggest that the site code be checked for problems, especially for subcalls. I’d start with Universitas Negeri Semarang, since the domain journal.unnes.ac.id shows as malware on nearly all that university’s OA journals–and also shows up as an infected call in quite a few other journals. Smaller domain problems include ojs.unud.ac.id at Universitas Udayana, ejournal.upi.edu at Universitas Pendidikan Indonesia and possibly others.

It’s possible that Malwarebytes is showing false positives, but I’m no malware expert, and since I’ve had malware problems directly traceable to journal sites in the past, I’m not willing to take the chance.

NOTE: These are not all new journals–indeed, most of them were in DOAJ before 2017 (102 appear to be new), but without malware.

You can download an Excel spreadsheet showing the title, URL, publisher and–in some cases–a note (“domain” means the URL up to but not including the first slash) as to the malware encountered. If you prefer, a comma-separated-values version is also available.

I hope someone’s able to clear these up; I look forward to rechecking the sites in late April or early May and finding few if any infected sites!

PS: About Romania

If you’ve read earlier progress reports on this year’s scan, you’ll know that Romania also had a number of malware issues: nine of the remaining 21 from the first 4,000 journals. Romania also had the most infected journals year, with 11. I am checking the remaining Romanian OA journals before going back to the alphabetical-by-publisher scan. Since there are only 137 more journals to check, I’d expect to finish this by next Tuesday (2/13) or so. If somebody out there would like to check on the Romanian journals–the probably-infected ones–send me email (waltcrawford@gmail) to that effect and I’ll provide a similar spreadsheet when it’s ready.

Note added 2/12/2017: I’ve done the Romanian scan, and it turns out there are only 11 infected journals in all. Here’s the discussion and a table.

I’m one-third of the way through the first pass.

Following up on the previous notes:

• The rate of malware occurrences has slowed down (but I can see that a whole bunch of journals from the most problematic nation are coming up soon).
• Indonesian OAstill seems to have a malware problem. At this point, of 120 Indonesian OA titles checked, 36 are flagged by Malwarebytes Pro as either being malicious or having malware-infected segments. That’s four more infections of 17 more journals.
• Other malware: 9 of 155 in Romania: 57 more journals, no new infections. 4 of 329 in Brazil: 149 more journals. one more infection. Two of 13 in Ukraine–four more journals, no more infections. Six countries with one malware case each: five more than previously.
• I’m hoping these will mostly resolve themselves on the second pass (or the third malware-only pass). I will not bypass Malwarebytes or Windows Defender protection. Period. Been there, done that, spent days fixing things.
• Note that there were only 67 journals flagged as malware in last year’s pass. I now have 57 flagged out of the first 3,600–with more than 7,100 left to go. The rate has gone down dramatically–but it’s far too high.

And a new note, only indirectly about the study. As I’d noted elsewhere, I seem to be having occasional system instability problems (although doing a full shutdown or restart once every day or so seems to be helping and only costs 2-3 minutes of lost time).

I thought I’d try LibreOffice instead of Excel, to see whether that would help with the problem; I’ve long suggested LibreOffice as a substitute for people who want Office functionality but either (a) despise Microsoft or (b) legitimately can’t afford $100 for Office.

After the lengthy process of updating to the latest version, I tried opening both spreadsheets and working with them. Unfortunately, what I found was:

• LibreOffice, faced with medium-sized spreadsheets (one’s about 400KB and currently has 3,600 rows of about 20 columns, with no formulas; the other has the same columns and lack of formulas but currently about 10,100 rows), was jerky: attempts to move around or adjust the window size were slow and erratic.
• One or two functions that I use regularly (e.g., column autowidth) just aren’t there.
• I found the integrated icon-based control bars confusing–but that’s because I’m now very comfortable with Office 2013’s contextual control bars.

After 10-15 minutes, always feeling that I was waiting for LO to catch up, I gave up.

(Notably, when I’d tried a similar experiment as a Word replacement, functions I consider necessary just weren’t there–but then, I create full-fledged books in Word, making more typographic control demands than the average user.)

I still think LibreOffice is a good no-cost alternative, especially for casual use. But, frankly, the price for Office seems like a bargain (no, I don’t use Office360: $100/year still doesn’t seem like a bargain.)