You have been blockedindefinitely from editing for running a bot script on this account without approval. If you think there are good reasons why you should be unblocked, you may appeal this block by adding below this notice the text {{unblock|reason=Your reason here ~~~~}}, but you should read the guide to appealing blocks first.
Under the bot policy, all automated scripts must be approved by the Bot Approvals Group to ensure that they perform safe and useful functions without stressing system resources.
This user's unblock request has been reviewed by an administrator, who declined the request. Other administrators may also review this block, but should not override the decision without good reason (see the blocking policy).
I edit on article space because I was testing error handling on Locked Pages, because I thought I was supposed to test before asking for permission. I'll go to BRFA now. VIAFbot (talk) 16:51, 17 August 2012 (UTC)[reply]
If you want to make any further unblock requests, please read the guide to appealing blocks first, then use the {{unblock}} template again. If you make too many unconvincing or disruptive unblock requests, you may be prevented from editing this page until your block has expired. Do not remove this unblock review while you are blocked.
Yes. Thanks for bringing this up. The irony is I was actually going to apply for BRFA today. Yesterday I made the edit to the chuck norris article because I was attempting to test error handling on locked pages. Anyway I'll move on to BRFA now. VIAFbot (talk) 16:24, 17 August 2012 (UTC)[reply]
This user's unblock request has been reviewed by an administrator, who accepted the request.
Note you should not be making edits such as this with the bot account. Use your regular user account instead. The bot account should only be used to make edits for approved tasks. Anomie⚔01:53, 14 September 2012 (UTC)[reply]
Thanks for making me aware of this. I was just making Userspace examples to test read from dewp. To clarify this bot never edits deWP only reads it. I've now blanked the pages. On another note, it's good to see the use of the TYP parameter, its a good thought for us in redesigning the Authority control template. Again, very sorry. Maximilianklein (talk) 11:43, 14 September 2012 (UTC)[reply]
An observation of which I am not sure if it really is problematic. However it might be worth the effort to sort this out, before proceeding ...
the connection between an article of en:Wikipedia and a VIAF cluster is described in wikilinksforbot.out – for this file it schould be true, that any VIAF cluster appears no more than once and any Wikipedia entry appears no more than once.
On the other hand, among the ~300000 VIAF clusters that have a link to en:Wikipedia, there are many thousand, where different clusters link to an identical article. Some examples:
In case of G. Wood the VIAF cluster selected in wikilinksforbot.out is the " wrong one".
I have no idea what percentage of VIAF links could be affected. 5% ... 10%? Just wanted to note this observation. Cheers -- Make (talk) 13:12, 15 September 2012 (UTC)[reply]
It is true that VIAF isn't perfect, but this is actually a great opportunity to improve VIAF, and that is part of the plan. The way the bot will handle this is that it will insert the first link that it comes across in the wikilinks.out file, then if it goes to edit that article again, it will detect a conflict and log it for human inspection [the bot userpage]. Then we will report these duplicates to VIAF as an organization who will correct it on the back end. They will then bring their clusters closer together. At that point we can develop a maintenance phase for VIAFbot to run on diffs supplied by VIAF.
Why does your bot move the stub template from the foot of the article to above the categories and below the defaultsort tag when adding this template? Please fix all the articles you've edited. LugnutsAnd the horse18:13, 4 October 2012 (UTC)[reply]
Ok, I think I see what's happening. I'm calling the pywikipediabot library to add the template below the text and above the categories. I think it's following a convention that there should never be templates below categories.
So can you explain to me exactly where to insert this template. Or even better, how you would algorithmically determine where to place the authority control template? I.e. Above categories, above interlanguage, above persondata etc. . Is there a complete list of templates of which to place it above? Maximilianklein (talk) 20:52, 4 October 2012 (UTC)[reply]
So above persondata, and if no persondata above defaultsort, and if no defaultsort above the first [[Category:...? That would be reasonably easy for the script to evaluate, I think. Andrew Gray (talk) 21:52, 4 October 2012 (UTC)[reply]
After a quick discussion on how best to fix this, I've now mass-reverted every VIAFbot change, allowing us to run the script again once the placement issues are sorted out. Andrew Gray (talk) 20:39, 4 October 2012 (UTC)[reply]
Resolved! (deo volente) The script now looks for persondata, and adds above it; if there's no persondata, it looks for defaultsort, and adds above that; if there's no defaultsort, it looks for the first category and adds it above that. If there are no categories either, it stops and logs the page for human examination. This will let us pick up uncategorised pages as a beneficial side-effect. Andrew Gray (talk) 22:48, 4 October 2012 (UTC)[reply]
Thanks for fixing the stub issue. I think there's another issue, but it looks like it's been picked up with the comments, above. The template is being placed between the header of the persondata template and the body of that template. See the bot edit on this article for example. LugnutsAnd the horse09:22, 6 October 2012 (UTC)[reply]
Well-spotted! I know exactly what it's doing - not understanding that the persondata comment is part of persondata and should be included. Hmm. It should be relatively easy to fix this by adding another exception rule - do you know if the header is always in the same format? Andrew Gray (talk) 10:52, 6 October 2012 (UTC)[reply]
There is a bot (don't recall which) which when it encounters a hidden comment on the tail of a line, such as
{{Persondata <!-- Metadata: see [[Wikipedia:Persondata]] -->
will split these, and put the comment on the line above. I do wish it wouldn't, because this divorces the comment from the line to which it pertains; and having been so divorced, well-meaning editors (bots or human) may see fit to insert other lines between them, further compounding the divorce.
For the specific case of Persondata, I've only found two "correct" forms: {{Persondata alone, and followed by the hidden comment as shown above. --Redrose64 (talk) 19:21, 6 October 2012 (UTC)[reply]
Ok, I've adjust to the algroithm to first search for <!-- Metadata: see [[Wikipedia:Persondata]] --> and place the AC template above it if it exists, as in this diff for example. Thanks for pointing out this bug. I'm going to run the bot on another 2,500 articles. Maximilianklein (talk) 20:51, 6 October 2012 (UTC)[reply]
VIAFbot adding second template to some articles[edit]
Thanks for spotting this; it seems divonly got created after we'd started planning the bot and so it wasn't something we'd thought about. It's very lightly used, though, so it should be possible to check over these easily - the only one affected at the moment seems to be d'Aubigné, which I've fixed by hand. I wonder if the best solution would be to merge the templates and have a divonly = yes display parameter? Andrew Gray (talk) 10:03, 5 October 2012 (UTC)[reply]
Maybe a parameter like "state=inline". This might allow for other display options later on. Template:Commons and Template:Commons inline are an example where they are left as separate templates. But I agree that combining into one template is probably the better choice for "Authority control". --Robert.Allen (talk) 19:52, 5 October 2012 (UTC)[reply]
I just discovered and fixed a case where the VIAFbot created an ugly mess in a major article, Michel de Montaigne. Hopefully you will have the bot's code fixed very quickly, as we do not want this kind of thing to spread. Here's the diff, so you can see just where VIAFbot went wrong. (Now I see that this is exactly the problem just reported by Allen3.)
Please take this as only constructive criticism. With professional training in library and information science (among many other things), I, for one, will be very interested to see how this project works out. I have also taken the liberty of adding a VIAF template manually to an article I have great interest in. But of course, as you know, we have to be very, very careful with bots. Like much modern surgery, their operations should be minimally invasive. Problems like this one are serious cause for concern. --Alan W (talk) 05:34, 9 October 2012 (UTC)[reply]
I see what's happened there! Interesting. The odds are this is likely to happen again, so I've stopped the bot until we get it fixed (which should be a small coding fix, but I'm not able to implement it from here). Andrew Gray (talk) 08:30, 9 October 2012 (UTC)[reply]
Actually I refactored the insertion algorithm to check for this now. There will be some lingering affected cases though, unfortunately. The newest algorithm check for the appearance of all the metadata comment, the persondata template, the defaultsort magic word, and the category links, and inserts above the the top-most occurence of all of those. Maximilianklein (talk) 15:31, 9 October 2012 (UTC)[reply]
Looking better now, thank you. I will continue to spot-check any VIAF insertions in pages on my watchlist and others I happen to notice, and I'll let you know if I see anything else peculiar. Regards, Alan W (talk) 01:10, 10 October 2012 (UTC)[reply]
We did think about this, but felt it was best to merge the two sets as we went along - it lets us do it in one go, rather than needing two passes, and means we can log and then fix the conflicts between the records before they go onto enwiki pages.
At least, that's the theory, it looks like it may not be working as perfectly as expected ;-) I've let Max know about this (he did the coding and has the logs), and we'll see if we can figure out what's going wrong. Andrew Gray (talk) 19:28, 9 October 2012 (UTC)[reply]
Ok, I've tracked down what happened in this case. The de.wp page was moved, and it rendered the interwiki-link a redirect. When the bot read the redirect page it had no Normdaten. This is actually an oversight, because I wrote a redirect-follower function in the bot when loading english pages. So I will adapt it for the deutsche. Kolja21, thanks for bringing this to my attention. To let you know, typically VIAFbot would have marked this page for manual checking, because the number it got and the de.wp disagreed, we check every page against de.wp and there are 13 cases we identified that are possible. Thanks again. Maximilianklein (talk) 22:25, 9 October 2012 (UTC)[reply]
+1. Any chance to make the lists of discovered conflicts available? I am sure some users of the Normdaten-group in de:wikipedia will help assess the conflicts. Especially types 6 and 11, which affect 3%-4% of pages touched. -- Make (talk) 14:05, 10 October 2012 (UTC)[reply]
Hey, I'm another user engaged in the German Normdaten activities. While doing some spot checks I found quickly another case of a wrong assignment: John Miller (linebacker). In the last few months we had worked with a list (with almost 7,000 entries in its original size) based on the combination of en-Wikipedia-VIAF matched pairs and de-en-wikilinking in order to add missing authority control numbers. Our (intellectually revised) log of mismatches has 139 entries (~ 2 %), a large percentage concerning sportsmen and women. So, based on that experience I'd recommend to double-check authority control data inserted into articles in the Category:Sports competitors. --Mai-Sachme (talk) 09:29, 10 October 2012 (UTC)[reply]
Ahh, good point. All the stats are being logged locally. So even though they are resetting each time I restart the bot once it finishes running I will post a full report, and make some pretty graphs in R. Maximilianklein (talk) 23:42, 9 October 2012 (UTC)[reply]
Our article about Rob Levin was recently merged into this one. Apparently the bot followed the redirect, adding an authority control template anyway. Is that really supposed to happen in this case? Shouldn't the bot ensure the article is in one of Category:Living people, Category:year deaths, Category:Year of death missing, or Category:Year of death unknown if the author is (or was) an individual? Also, can you explain how the VIAF web site ended up with the Wikipedia links in the first place? (I'm wondering whether both Wikipedia and VIAF are referring to the same Rob Levin born in 1955.) Thanks, PleaseStand (talk) 03:26, 10 October 2012 (UTC)[reply]
The bot follows redirects because (in most cases) they're simply renames, but a check for "looks like a person" is definitely worth doing - I know we talked about it at some point, but it doesn't seem to have made it into the final version. I've stopped the bot for now and will try to get it running again this evening with some form of check in it.
VIAF use Wikipedia data for their clustering algorithm, which is why we have the existing set of links. In this case, it looks like an error anyway - he's probably not the photographer! Andrew Gray (talk) 09:24, 10 October 2012 (UTC)[reply]
Thanks for spotting this. It's important to understand that VIAFbot is using data that did check the Persondata and categories to match. The reason why we're following redirects is because there are a lot of cases like "John Smith (Footballer)" -> "John Smith (Goalkeeper)". The reason we can't just assume that those pages will be in Category:Living people, is that we are actually running the list of "Personal Names" which includes Fictional Characters who are not Living People. This case is unfortunate, and I'm really pleased that you could spot it and clean it up by hand. Maximilianklein (talk) 17:17, 10 October 2012 (UTC)[reply]
Hi — the bot just added a VIAF for Hans Lewy that looks very strange to me: it combines information about what looks to me like two different people, Hans Lewy (a mathematician) and someone else named "Yochanan Lewy" (a historian of Judaism). The biographical references in our article (e.g. an obit in the New York Times) say nothing about an identity between these two people. I realize the bot is just making links, but in this case the link seems half-wrong to me. Do you have any suggestions for how to get this resolved, either by cleaning up the VIAF entry or finding proper evidence of identity that we can use in our article? —David Eppstein (talk) 05:41, 10 October 2012 (UTC)[reply]
I can see no evidence that the authority record here is right - but have emailed the English MP's office to ask for confirmation that he is not the author of the American book about an American charity. If I get that confirmation, I'll be back here to ask how we get the VIAF record amended, as it currently includes the book authorship and has a link to the (previous title of the) Wikipedia article! It's not unlikely that two people of this name were born in the same year. PamD07:10, 10 October 2012 (UTC)[reply]
Thank you for your close attention. It's been reported on the newly created error page Wikipedia:VIAF/errors, if you see any more please report it there and we'll either correct it, or if necessary report it to VIAF.org for corrections. Maximilianklein (talk) 17:19, 10 October 2012 (UTC)[reply]
There is a known error rate, we understand. At RfC we estimated the error rate, based on a random sample, to be 2% and found consensus to run with the 2% error rate. The bot is stopped at the moment, I will run another random sample of edits and assess the error rate and report it before continuing. Maximilianklein (talk)
Thanks for reporting, if it's an error with the bot, I will fix that. If its an error in the article that VIAF.org is pointing to on their site, then if you log it on WP:VIAF/errors we'll report the error to VIAF so it can get fixed at the source. Maximilianklein (talk) 17:27, 10 October 2012 (UTC)[reply]
That one is weird - we talked a bit about multiple people, but the idea of control records for fictional authors escaped me. Hmm. I think I might try generating a report for the intersection of {{authority control}} and the various subcats of "fictional people" and see what else we have.
I don't actually know what's best here - on the one hand, they're definitely not people, but on the other hand, they are "people-like identities". How should we handle cases like Junius (VIAF), I wonder, where the "authorship" undeniably exists but the correlation to a real person is unknown? Andrew Gray (talk) 09:02, 10 October 2012 (UTC)[reply]
I'm not sure how it should be handled, so long as it's handled consistently. It's tempting it say link it to persondata or something similar. Stuartyeates (talk) 09:20, 10 October 2012 (UTC)[reply]
Okay, here's my suggestion. After thinking about it for a bit, I feel, in general, we should consistently aim to keep the record for "nominal" people as well as real ones, simply because otherwise it's very difficult to draw the line - there's a continuum between "real authors" on the one hand, through pseudonyms with no person attatched (Junius), nominal-but-nonexistent authors (Carolyn Keene), mythical authors (Homer) and outright fictional ones from within the fictional work (Mutzenbacher). Cases as extreme Mutzenbacher are rare - most cataloguers will just attribute these to the real authors - but some persist.
In this particular case, though, I think it looks weird mainly because the article is mainly writing about the book. In cases like this - where the article is presented as something other than a biography or pseudo-biography - suppressing the authority control field is the best option editorially; it avoids confusing the readers. It also makes sense for cases like the Freenode one above!
Getting the script to identify cases like this is tricky, though. I had wondered if we could do this by using things like subcats of Category:Novels, but experiments with catscan suggest that's >95% false positives. Andrew Gray (talk) 21:17, 10 October 2012 (UTC)[reply]
It should be looking for normdaten as standard - I'll flag this one as a bug. Thanks for spotting it. Normdaten is (thankfully) the only redirect template, so what I'll do is, rather than stop the bot pending a fix, run a report looking for any cases with both and fix them manually. Andrew Gray (talk) 14:16, 11 October 2012 (UTC)[reply]
Wow VIAFbot! You are improving biographical articles which I watch by adding the authority control template! I like what you do a lot. If Wikilove had pictures of cups of oil then that might be a more appropriate way to thank a robot than by giving this cup of coffee, but this is what I have. Blue Rasberry (talk)15:50, 11 October 2012 (UTC)[reply]
there seem to be quite a number of what I would call pseudo conflicts: When VIAF clusters are merged, old identifiers (URIs) are not invalidated but instead transformed into a redirect. This is explained at [2] Thus different VIAF-IDs can still identify one and the same cluster. Example: Current VIAF cluster for Arthur Adams (comics)ishttp://viaf.org/viaf/100977426/ however http://viaf.org/viaf/90725635/ (from de:Normdaten) redirects to this cluster as well.
Interesting! I see what happened - it's programmed to look for various elements (categories, persondata, etc) and place the template above the top-most one. It saw [[Category and thought it was an actual category link; I'll get it changed to look for [[Category:! Thanks for spotting it. Andrew Gray (talk) 20:09, 12 October 2012 (UTC)[reply]
Yup, sorry I was copying pages from WP into userspace to test my bot in a noncritical place, and the categories came along with the wikitext which was copied. I'll fix it now. Maximilianklein (talk) 15:23, 18 October 2012 (UTC)[reply]
This occured because the bot was set to follow redirects for pages, and I see now that some pages redirect to DABs now. I updated VIAFbot now to check for DAB pages, so this shouldn't happen anymore. There isn't an easy way to check for which DABs this has already occured on, but thanks for alerting me. We are less than 25% of the way through, so it's good that we caught this earlier rather than later. Maximilianklein (talk) 00:59, 22 October 2012 (UTC)[reply]
Rethis edit the resulting link page at the foot of the article does not seem useful. What is its purpose? Is there a discussion thread on this? The link wrongly has a US flag next to his birth. Eldumpo (talk) 22:43, 8 November 2012 (UTC)[reply]
Hi, the bot apparently does not recognize the Normdaten template and so redundantly adds an authority control template where it already exists: evidence (the added template is also worse than the existing one). Does this happen often, does the bot have to run over these instances again or do we have to fix the double templates manually? Hekerui (talk) 23:24, 8 November 2012 (UTC)[reply]
Thanks for bringing this up. There are only about 50 Normdaten templates in use and they redirect to Authority contorl. I'm going to just quickly convert them. Thanks for your attention to detail. Maximilianklein (talk) 17:53, 9 November 2012 (UTC)[reply]
I went through all these yesterday and could only find Blazek. I've double-checked now, and turned up a couple of odd cases with no displayed elements (I think the template's been copied from German) which I'll look at later, but I can't track any more cases of double-use. Andrew Gray (talk) 09:58, 18 December 2012 (UTC)[reply]
Yes the template is indeed copied from the German, as are the awful capital letter parameter names for "pesondata" (the clue is in the agglutinative names for the templates, which we avoid). RichFarmbrough, 23:03, 18 December 2012 (UTC).[reply]
Hi there. I noticed that apparently now Wikipedia data is also an (experimental) source for VIAF. See Björn Gunnlaugsson and his VIAF entry. While I think that is great (and I also see that this is still experimental), I see several problems there:
The name somehow got converted to all lowercase.
I'll attribute the missing Umlaut in『Björn』to a poor {{Persondata}} template in the article (corrected now).
Whatever heuristics is used to generate "selected titles", it goes wrong too often. Of the 10 entries from Wikipedia, only the first two are correct. Two more (Njóla and De mensura et delineatione Islandiae) are missing. A 20% hit rate is probably too low to be useful. (I'll readily grant that figuring out such information from free-text Wikipedia entries is hard.)
Mind you, I'm not complaining. I think it's great that WP data is going into VIAF. But I see a dire need to improve that process. Lupo06:47, 9 November 2012 (UTC)[reply]
I'm in contact with the VIAF engineers. What's interesting is that all the WP information in VIAF was always used to help link the national authority records, it was just simply hidden from the user. Now the decision is to make that public. As for the lower case issue, they know its an issue and a fix is in the works. Thanks for bringing this up though. Maximilianklein (talk) 17:56, 9 November 2012 (UTC)[reply]
I can't believe I made that grammatical mistake, well it's fixed now, and @Stuartyeates, if there are unforseen conflict types like the ones you stumble on then let's just be bold and create new categories for them. When I deliver the list to VIAF there will be a human correcting these so error categorization is not terrifically important. Maximilianklein (talk) 00:40, 21 November 2012 (UTC)[reply]
Hi -
I don't know whether this constitutes an error or not, since I can't find a succinct explanation of what you are trying to accomplish. However, these results seem inconsistent and confusing so I thought I'd point them out. Looking at 3 entries you recently did,
29524192 lists mostly works by Stefansson but also Stefansson and the Canadian Arctic which is about him (there are also a couple I don't recognize and haven't attempted to research).
104465315 lists works from the bibliography of the WP article on Richardson Clover, about Clover rather than by him, and also 2 entries that apparently reflect italicized terms in the WP article rather than books, inside passage and uss susquehana 1850. The book cover illustration shows and links to a multi-volume treatise which mentions him in one sentence; does not list the relevant page.
24252928 includes works by Charles Wesley Leffingwell and includes multiple variants of the same title. This entry does not include works about Leffingwell cited in the WP bibliography which I presume reflects your intention.
If you are going to automate the extraction of bibliographic information from WP, you need to be aware that bibliographic formats and article formats both vary widely, and that italics are used for purposes other than titles. Dankarl (talk) 16:43, 13 November 2012 (UTC)[reply]
Thank you for raising this issue. In fact the process of VIAFbot putting in viaf.org entries, and the process of scraping pages for selected works are unrelated. So this isn't really within the scope of this bot. However I know the team that is working on the selected titles scraper, and I will forward your message, since that bot works on a database dump and has no account. Maximilianklein (talk) 17:36, 13 November 2012 (UTC)[reply]
FWIW 100436106 also contains a mix of by, about and simply italicized items. Your database-scraping cohorts might do well to open a page here for similar reports. I won't bug you further with these. Dankarl (talk) 18:37, 15 November 2012 (UTC)[reply]
That's because there are many different authority control standards, not only does VIAF offer this service, but so does the US Library of Congress (LCCN) and German National Library (PND) etc. However VIAF is a "bridge identifier" which means that ideally if you had the VIAF ID of person you could find their LCCN and PND, by searching at viaf.org . The matching is still not 100% complete or accurate, but that's the theory. Given that theory I am attempting only to place the VIAF numbers, and on commons people have different opinions. Maximilianklein (talk) 17:42, 19 November 2012 (UTC)[reply]
I know I just thanked you a month ago but it seems like every day since then you have added more authority control templates to more biographies which I watch. VIAFbot, your work is so important. Blue Rasberry (talk)14:47, 19 November 2012 (UTC)[reply]
Hello there, you just edited something on the Garbo page which I don't understand. Can you explain what you did? People make smaill adjustments which I don't understand so, out of curiosity, I want to learn this stuff! You can just answer, if you wish, on my talk p. thanks,--Classicfilmbuff (talk) 00:53, 26 November 2012 (UTC)[reply]
you are putting authority control in an article that you have made no contributions to such as the J.B. Vuillaume article.
there is no reason for that. Milliot (talk) 07:46, 26 November 2012 (UTC)[reply]
I don't think it's necessary to revert the Authority control template that VIAFbot put there. Over .25 million Wikipedia pages now use the Authority control template as disambiguation device. The bot not contributing to the article otherwise isn't critera for removal in my opinion, it's just making an incremental upgrade to the encyclopedia. Maximilianklein (talk) 22:08, 26 November 2012 (UTC)[reply]
I think initially somebody had created a page - Josef Israels - a Dutch painter. However the correct way of spelling his name is Jozef with a Z rather then S ! So somebody must have changed the s to a Z recently ?! I have mentioned on the page - George Paul Chalmers that he painted the portrait of Jozef Israels but then it appears in red and not in blue whereas if I incorrectly type it with a s - it does appear in blue ??
I am not a very literate computer person, so I do not know how to solve this technical computer problem !
Glemmens1940 (talk) 12:55, 21 January 2013 (UTC)[reply]
Hello, I find that this VIAF stuff is rather intrusive -- it has zero interest for the average reader, yet sits on the article page rather than the Talk page, which would seem more sensible. Would it be ok if, in the articles I'm seriously involved with, I moved the VIAF to the talk page instead? Thanks, Opus33 (talk) 20:26, 27 January 2013 (UTC)[reply]
I'm not sure that it holds zero interest for the average reader - VIAFbot went through an RFC and found consensus from the community as a whole, to be on the article page and not the user page. Technically it would work if VIAF IDs were on the user page and not the article page, but they are on the article page for a reason, the community thinks they are useful. So I would ask you not to move the VIAF IDs please. Maximilianklein (talk) 00:16, 28 January 2013 (UTC)[reply]
@Roseohioresident: If the core problem is with Wikipedia pointing to the wrong cluster, then fix it on wiki, and VIAF will notice the change. If the problem is that the VIAF cluster confuses two people then you can send a message with the problem to oclcresearchoclc.org. Maximilianklein (talk) 19:55, 3 September 2013 (UTC)[reply]
VIAF 84033063 stands for Wilfred Ernest Lytton "Will" Day (1873-1936), a British film historian. I've deleted the link in Wikidata. --Kolja21 (talk) 18:48, 9 September 2013 (UTC)[reply]