User talk:Materialscientist

Please leave your message at the bottom of this page. I shall reply wherever you prefer (my usual habit is to reply here if the answer is short). If I replied on your talk page it means I am watching it, and there is no need to add {{talkback}} template or quote the previous message.

Impact factors and infoboxes

Hi, I noticed that you updated several IFs, but then also removed this from the body of the text, together with what often is the only independent reference that an article has. Please don't do this. An infobox provides a rapid overview of what is in the article. And while we don't repeat basic bibliographic info (ISSN, OCLC, etc), it is no problem to say something about an IF. In addition, the reference avoids the article being tagged for having no independent references or no references at all. As for ranking info, that can easily be updated at the same time that an IF is updated and there is no need to remove that info either. Thanks. --Guillaume2303 (talk) 06:59, 29 June 2012 (UTC)

In practice, having a duplicate in the body is bad because it is easy to miss it during updates. I don't see a difference between the impact factor and other infobox parameters - they are all assigned by some authority. In other words, the reference to Journal Citation Reports is a dummy, because there can be no other source for the impact factors (in reality impact factors are issued by ISI - no-one may or can "calculate" them), and it is not possible to directly link to that database. Materialscientist (talk) 07:12, 29 June 2012 (UTC)

Sorry, but I disagree. There's a fundamental difference between an ISSN and an IF. The former is just assigned in a kind of arbitrary way (it says something about which publisher or something like that, but otherwise doesn't convey any info), but the IF is something that is calculated by a reliable source. That you or I cannot independently redo those calculations is immaterial. And even though one cannot directly link to the database, neither can we do that for a print source and we can still use that as a reference and the JCR certainly is a reliable source. Putting in those references has been a lot of work and provides needed sourcing to journal articles, so I'd really appreciate if you would not remove them. Thanks! --Guillaume2303 (talk) 09:38, 29 June 2012 (UTC)

Impact factors are assigned by the ISI and no-one else (yes, sort of calculated, but in a non-transparent way). Thus adding a ref "Journal Citation Reports 2011" is not "a lot of work"; it could be done by a bot and is simply redundant. Materialscientist (talk) 09:46, 29 June 2012 (UTC)

I find that a strange reasoning. So each time there's only one possible source for something, we don't cite it? And the refs are usually much more detailed than "JCR 2011" and go something like: "<ref name=WoS>{{cite book |year=2012 |chapter=JOURNALNAME |title=2010 [[Journal Citation Reports]] |publisher=[[Thomson Reuters]] |edition=Science |accessdate=2012-06-29 |series=[[Web of Science]] |postscript=.}}</ref>". I don't see what is redundant about this. --Guillaume2303 (talk) 10:04, 29 June 2012 (UTC)

When there is only one source, and the data are integrated into the infobox, that source can be simply included into the infobox or impact factor wikilink. The ref above is incorrect. Journal Citation Reports is not a book. It is a regularly updated electronic database. It has no sections/chapters/editions - instead, specific datasets can be extracted from it through search menus provided by the interface. Thus the output table/page depends on the input. Materialscientist (talk) 10:12, 29 June 2012 (UTC)

Sorry, but that is all wrong. The JCR has been around for a long time and for most of it existence, it was a print publication. One volume (or book, or whatever you want to call it) was (and still is) produced annually. It contains for each included journal a section that present the different citation metrics (impact factor, immediacy index, etc), together with details on their calculation. For different fields (Neuroscience, Archeology, etc) it has a chapter that presents tables with rankings based on these different metrics, containing only the journals that are included in that specific category. I don't know whether ISI still produces the print edition, but the JCR still is organized in the same way. And even though it now has online search possibilities that permit you to, say, list all journals whose titles start with "A", the journals are still basically organized by categories. So, yes, I think a category is a "chapter" of the "book" JCR. This goes even more so for the individual "chapters" about the journals, because there you cannot modify the content by formulating your query different. For each journal the same information is displayed in the same way. So its not like PubMed, where the list of articles depend on which keywords you enter. It really is a book, just that nowadays it is in electronic form, much like encyclopedias for centuries were books and only recently have gone online (and you wouldn't call WP a "database", now would you?) By the way, having said all this, I don't think things would be much different if JCR really was a database, it would still be a reliable source and citable. --Guillaume2303 (talk) 20:53, 29 June 2012 (UTC)

<indent>Sorry, I can't make time for a proper answer. Indeed, the entries for individual journals are chapter-like, but impact factors can be retrieved as various tables through search queries - this is what I meant by database. A reference "Journal Citation Reports XXX" will suffice, as indeed, JCR is absolutely reliable, but I believe this should be integrated into the infobox (e.g. as a link to the impact-year). I'll have a look at the formulas they give on monday, but there is an obvious trick to them: while the formulas are trivial math, the crucial variable (number of cites) is basically defined by the ISI (they've got various "correction" procedures, and simply taking citations by year from WoS won't give the impact factors). Regards. Materialscientist (talk) 12:37, 30 June 2012 (UTC)

The trick is not all that difficult. For years I was EIC of a journal myself and every year I was able to predict our IF within 1 or 2 decimal points. The corrections are made by looking at citations with typos in them. These don't get assigned to the correct article in the Science Citation Index database. So you first use "search" to get all "correct" citations (you can use the citation analysis tool for that) and then use the "cited reference search" and only look at the unlinked references and manually count the numbers of citations. Then you combine the two searches and you come pretty close to what the JCR gives. Discrepancies occur if there are citations with typos in the journal name (ISI finds them, we can, too, but it's a lot of work) or when counting the number of "citeable articles" in a given year (sometimes it's debatable whether an item is going to count as "editorial" or "article"). If there are a lot of these cases, the discrepancy between estimated and "real" IF can become quite large, of course. However calculated, the IF is a unique value assigned by a reliable source and should be referenced. And although the references I have placed in many articles all look pretty much the same (but so do references to, say Encyclopedia Britannica), they are not identical. --Guillaume2303 (talk) 17:08, 30 June 2012 (UTC)

JCR is the customary source, but they are not "absolutely reliable". There is a good deal of possible variation over how impact factors should be calculated. The definition is due to Eugene Garfield specifically, not ISI in general, and ISI calculates them on the basis of its particular range of journals and assumptions. A. The basic assumption for the classical impact factor is its 2-year basis for citations. (thus the wording impact, not influence--it measure the initial effect,) This fits very well fields like molecular biology, but not fields with longer citation patterns. So in fact, ISI also calculates a 5-uear impact factor. It is also possible to use their data and calculate for any arbitrary time basis--the explanatory material in JCR explains in detail how to do so. B.The factor is not a reflection of the total universe of potential citing journals in the world, but of those particular journals that have been chosen for the JCR set. Citations appearing elsewhere do not affect the JCR facto . Thus the IFs are gross underestimates for journals that may be included, but which are primarily cited by other journals the ISI does not cover, as for most third world journals. Though a Chinese language journal may be included, most of the citation on it will be from other Chinese journals that are not included. Therefore, both China and Japan also calculate national-based impact factors for their journals. (The ISI IF however does have validity, if it is taken as measuring the immediate influence upon the world wide core scientific community. )C. It is not necessarily obvious what is a citable item. The basis does not include items in journals such as editorial introductions and announcements which would not ordinarily be cited anywhere--this can be quite large in some journals. If they are not totally excluded, the impact factor will be lower than it would be if a narrow definition is chosen. D. journals Published in multiple editions need the citations combined if the scientific contents are the same. ISI attempts to do this, but their completeness varies. E The IF of a journal is affected by the proportion of review papers it publishes, because these normally have considerably higher citations. The pure review journals thus have very high IFs, but many journals have a few such papers. Again, ISI gives directions for eliminating this variation, though few people do so. (the most prevalent fraudulent manipulation of IFs is not self-citation --whose effect can, btw, also be measured--, but the addition of a few commissioned review papers to an otherwise primary journal.) F.There are also the errors that Guillaume mentions.

Additionally, I consider JCR impact factors when given as raw numbers without consideration of subject to be utterly meaningless. They were never intended as numbers valid across the entire range of subjects, but only for comparing journals of a certain type within a particular subject having uniform citation pattens. There's been an immense amount of work by various people trying to normalize this: there is no totally accepted standard, the most widely used is the so-called Eigenfactors they & Scopus publish. I always give the data by saying "the Impact factor for 20xx, as calculated by JCR, is x.xx, placing the journal nth within the m journals in the field of whatever. The citation needs to be to the database as a whole. (I have been customarily assuming the 2-year IF, which is relevant to most of experimental biology, but probably I should start saying so specifically. I think it wrong to give then umber without citation. It is a number calculated in a particular way, not an observable fact of nature. (I should mention that Garfield & ISI are well aware of all this--his collected works, online, discuss all of the variation and interpretation problems. He has always written against naïve use of the data. There have been universities --and countries--saying that the only papers that count for funding are those published in journals of IF about a certain value (one value I've seen is 1.5) for all fields. This is nonsense. ).

For WP, I therefore think that the numbers should be given , cited, in text. That we also give them in the infoboxes is in my opinion pandering to those who use the numbers without understanding. That opinion has not been generally accepted here, so I do add them to the boxes when I write a journal article. Infoboxes need to be consistent. I hope the Wikidata project will lead to greater uniformity within and between the different wikipedias. DGG ( talk ) 16:56, 1 July 2012 (UTC)

PoO2

How's the polonium dioxide GAN? Double sharp (talk) 09:15, 30 June 2012 (UTC)

Uranium

Regarding fallout from atomic bombs, I think that the current version of the article is somewhat misleading. I saw your edit summary. While I am aware of a few pure uranium bombs being detonated, the vast majority of bombs used Pu as the main fission fuel. I recall reading that about 20 % of the fission occurs in the uranium tamper so the vast majority of the fission is of Pu. While Pu is a daughter of U-239, I think that the text is misleading. Of the big H-bombs I do not think that any went for the highest power design which would be a fission primary followed by a fusion secondary inside a U-238 tube to give a fission-fusion-fission design.Dr Mark Foreman (talk) 22:51, 30 June 2012 (UTC)

As I understood that section upon a brief look yesterday, its purpose was just to say that measurable amounts of uranium in nature originated from nuclear tests and nuclear accidents. This seems factual. Uranium isotopes were not a major product, and this can be briefly mentioned, but is not the topic of that article. Regards. Materialscientist (talk) 00:27, 1 July 2012 (UTC)

Admin?

Are you? If so, should Philcha's talk page be protected too as he's gone? Thanks. Atum World There's an Acadia for that too! 03:37, 2 July 2012 (UTC)

Vanriperandnies (talk · contribs)

Hi, Materialscientist. I had blocked this user with a username soft block immediately antcedent to your message on their talk. If you would like me to unblock, please leave a message here. I won't get in your way if you want to take another path. Regards Tiderolls 04:57, 2 July 2012 (UTC)

Glad to see you around. My current habits are these: (i) I am fine with extension of any of my blocks - no need to ask (unless this is a regular contributor, but I very rarely block those). (ii) When I meet inappropriate username, I warn and/or wait to see their intentions, and then either hardblock or softblock. I don't softblock (allowing account creation) right away because this may not resolve potential problems. (This is why I warned Vanriperandnies). Cheers. Materialscientist (talk) 05:17, 2 July 2012 (UTC)

I will unblock and leave a note of my own. Tiderolls 05:20, 2 July 2012 (UTC)

Material Scientist is a dick

Hi Material Scientist, I think you're a dick. Why? Because you complained to me regarding my John Bolton edits. You just don't get it---or perhaps you're a teabagger? Neo-con? Bush lover? Regardless, it was my patriotic duty to edit wacko Bolton's biography because, simply put he's a ranting, raging lunatic. He among others has caused immense damage to the USA both at home and abroad.

I hate goody-two shoes assholes like you. Clueless wanker. I'll be sure to edit Bolton's page again....mwahahahahahahah

Gerald Anderson

I hope you won't mind, but I need help on taking care of some fanboy who keeps on pushing his BLPvio edits, despite myself trying to clean things up to a more decent state. --Blake Gripling (talk) 10:45, 2 July 2012 (UTC)