User talk:Stephan Schulz

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Greetings[edit]

Hi all!

I'll answer all messages left on this page here, so that a possible discussion is kept in context. Watch this if you are waiting for an answer.

--Stephan Schulz

Archive

Archives


2004-12-13 to 2008-04-15
2008-04-15 to 2009-01-22
2009-01-22 to 2009-09-01
2009-09-02 to 2010-04-14
2010-04-14 to 2011-06-16
2011-06-17 to 2012-08-02
2012-08-03 to 2013-06-21
2013-06-22 to 2015-12-22


Useful links (courtesy Angela 02:29, Oct 31, 2003 (UTC))[edit]


Redirect[edit]

Hi Setphan, I am afraid that if you won't take the action now, it won't happen in the coming months at all: Should you redirect "Data Serialization Languages" to "Serialization? If in computing, these terms are enough synonymous, it can be nice.

Ben-Yeudith (talk) 01:42, 11 January 2016 (UTC)

Done, albeit for the singular data serialization language. But every user can create redirects - it's not a special right for admins. See Wikipedia:Redirect. --Stephan Schulz (talk) 07:50, 11 January 2016 (UTC)

Wheat[edit]

Wheat, yes, wheat. If you pick a species that has trouble with heat, you will see that. But is that a good way to evaluate the overall situation? I've come to your talk page to ask a question -- is the goal of this type of discussion to reach truth, or to make debating points? (actually there is a third possibility -- you may be genuinely unaware that it's not a representative example). I've come to your page because I believe such a question is better asked in a less-public forum. I realize your page is still public, which is unfortunate; I would ask it in a fully private way if I could do so. Feel more than free to take your time in answering, or for that matter, to delete this on sight. Best wishes. Really. CometEncke (talk) 15:04, 21 January 2016 (UTC)

Hi CometEncke! Let me first point out that I linked to two studies (the first two that sprang to my eye as applicable), and that only the first concentrates on wheat, while the second looked at about a dozen different crops. What I try to say is that this is a complex, multivariate problem and that the simple answer is likely to be wrong. I'm usually always interested in getting to the bottom of the subject matter (some say to a fault), but I've come to realise that I cannot do that with all of reality (despite the hubris expressed by the Bacon quote on my front page). So unless it's really within my field or I'm extraordinarily interested, I apply some heuristics. One such heuristic is that if someone says that something is "obvious" or even "blindingly obvious", I take a short track to Google Scholar and check some related papers. If that shows me that the problem is indeed complex and multifactorial, I assume this is a case for H.L. Mencken ("...neat, plausible, and wrong"). In this case I think that the nuanced, conservative discussion by the IPCC is more justified than "it works in my greenhouse for cucumbers, therefore CO2 will save us from world hunger" (sorry for the hyperbolic summary - I trust you get my point). If you want to convince me otherwise, you would either need to make an incredibly clear and short and convincing argument, or get your opinion published in a serious peer-reviewed venue, so that I know that you convinced real experts of the validity of your argument.
If you have a real need for private conversation, my email is enabled. But I normally prefer to keep Wikipedia discussions in the open - after all, making knowledge accessible is the whole point of the project. --Stephan Schulz (talk) 16:19, 21 January 2016 (UTC)
Well, OK, but as any farmer from a hot region will tell you, wheat is not for the hottest weather. Rice and millet do much better; maize can also be an option, though extra CO2 doesn't help it much. If your goal is to come up with crops that do poorly in hot weather, even with extra CO2, you can find them. But from the standpoint of food production, surely the more relevant question is not whether there exist crops that do poorly under those conditions, but rather, whether there exist crops that do well under those conditions. And the answer to that is a resounding "yes", as a little googling will tell you, or even just checking the population figures for South Asia. If you follow the evidence wherever it leads, regardless of where you may want it to lead, you will find that multiple, independent strands of evidence all point to the same conclusion, including the CO2 data themselves (the trend of the May-October drop, that is). And yes, it will be blindingly obvious, screaming at you like the evidence in a murder case where the jury got it wrong. CometEncke (talk) 17:55, 21 January 2016 (UTC)
CometEncke, congratulations on your blindingly obvious thoughts – however, for this reasoning to appear in Wikipedia, it must first be published in a reliable third-party source. A scientific journal is the best place to get credence. . . dave souza, talk 18:11, 21 January 2016 (UTC)
I have the feeling that you only read about half of what I wrote. Anyways, even stipulating that you are right that there are some crops that will do better, it's far from trivial that it's possible to replace existing agricultural systems - you need not just suitable climate (and remember that that will keep changing for a while), you also need skills, seeds, suitable soil, tools, and markets, to name just a few. --Stephan Schulz (talk) 19:00, 21 January 2016 (UTC)

Where's the Bacon? My search-fu didn't find it, am hoping this was a reference to the title page of On the Origin of Species, but any Baconian epigram will be of interest. . . dave souza, talk 18:14, 21 January 2016 (UTC)

I have taken all knowledge to be my province, which according to q: FrancisBacon is from a letter to William Cecil, 1st Baron Burghley. My pretentious user box is at User:Stephan_Schulz/knowledge. --Stephan Schulz (talk) 18:53, 21 January 2016 (UTC)
Thanks! Very modestly placed in the userbox, didn't know about that one. . dave souza, talk 19:15, 21 January 2016 (UTC)
"I have the feeling that you only read only half of what I wrote." In one case, you are right. After I looked at your "wheat" paper, I felt it was sufficiently irrelevant that I didn't go to the second one. Interestingly, this whole discussion got started because I felt pretty much the same thing and said so on the article talk page, not only in relation to you. So perhaps in that sense we are alike. I agree with you that our discussion has gone beyond the level of what can be included in Wikipedia, which is an additional reason to take it here, as opposed to the article page. As to your question of whether the agricultural system can keep up, I would refer you to a graph of worldwide grain harvests. I am curious whether you still think your "wheat" argument is evidence against my assertions about where agriculture is likely to head in a high-carbon world? DS -- welcome to the party. The scientific journals do talk about this all the time -- here is two reviews of a whole lot of literature, [1][2], which, according to google, have been cited 400 and 600 times, respectively. A whole lot more out there. The evidence is truly overwhelming. As this user astutely notes, facts do not cease to exist because they are ignored. Fortunately such is the case for food production and will remain so no matter how many people mock me for saying "blindingly obvious". CometEncke (talk) 12:33, 22 January 2016 (UTC)
Now I have not only the feeling that you read only half of what I wrote, but that you also read things I never wrote, and don't really read things you suggest we should read ;-). I have not made "a wheat argument" - I have made a plausibility check on you claim that it is "blindingly obvious" what the effect of higher CO2 on agriculture is, and found that it is very much not obvious. Indeed, the first source (in New Phytologist) you offered above very much said so. The second (the Oecologica paper) has very little predictive relevance for real-life open-field agriculture, as it is not a literature review, but a description of one experiment in a glass house with otherwise controlled conditions. Even then the paper concludes that "these data show that plant responses to elevated atmospheric partial pressure of CO2 depend on complex of partially compensatory processes which are not readily predictable". So I'm at a loss to understand where your "blindingly obvious" is coming from. As for the graph: You do understand that grain harvest are not primarily influenced by CO2, but by new cultivars and by the increased use of energy- and nutrient-intensive farming techniques, many of which are unlikely to be sustainable in the long term. Indeed, I find it a bit ironic that that the first hit Google gives me for "graph of worldwide grain harvests" starts with ""Global Grain Stocks Drop Dangerously Low as 2012 Consumption Exceeded Production... The drop was largely because of droughts that devastated several major crops—namely corn in the United States (the world’s largest crop) and wheat in Russia, Kazakhstan, Ukraine, and Australia". --Stephan Schulz (talk) 13:11, 22 January 2016 (UTC)
"Require further study". Sure, there is a lot we still don't know, and this was even more the case in the past. There is plenty of debate about all sorts of details. I'm still curious if you still believe your wheat reference is significant. If so, why? If not, why not say so? You may have noticed in this discussion that I have had no hesitation agreeing with you on certain questions when I thought you were correct. I guess I'm challenging you to do the same, or, if you can't, to explain it. Such action would give me confidence that a search for truth is more important than plausibility attacks, and would further give me confidence that moving on to other issues has value. CometEncke (talk) 07:19, 23 January 2016 (UTC)
Sorry, but my perception is not that you agree on certain questions, but that you are moving the goal posts. Remember, this started out as a discussion over at talk:Marco Rubio about the question if a number of sources, one of which headlines (!) "Marco Rubio says human activity isn't causing climate change" are enough to support the claim that "Rubio disputes the scientific understanding of climate change, arguing that human activity does not play a major role in global warming". Then you made claim about the IPCC (which Boris has refuted over at User_talk:MastCell#Rubio_and_climate - let me refute it here again, more explicitly: "Evidence since AR4 confirms the stimulatory effects of carbon dioxide (CO2) in most cases and the damaging effects of elevated tropospheric ozone (O3) on crop yields (high confidence). Experimental and modelling evidence indicates that interactions between CO2 and O3, mean temperature and extremes, water, and nitrogen are nonlinear and difficult to predict (medium confidence)" (executive summary, page 488, emphasis mine))) and about the "blinding obviousness" of the influence of CO2 on agriculture. Now we are discussing if wheat, a major staple crop and the major source of plant protein is a good example for discussing the claim that the effect of increase atmospheric CO2 on agriculture is "blindingly obvious" - apparently because wheat is more sensitive to heat than some other crops. So let me state it here: Yes, I think this is a relevant example for the claim that the situation is not blindingly obvious, but complex. If you are looking for something that we probably agree on: I agree that an increase in CO2 from the base level to a moderately increased level alone increases primary plant productivity for most plants in situations where growths is not limited by the unavailability other resources. But that is something the IPCC acknowledges as well. It just doesn't stop there. --Stephan Schulz (talk) 10:09, 23 January 2016 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Thank you for addressing my question. And sorry for responding slowly -- real life had me off wiki for a few days. Now for the question about the overall effect of CO2 rise on plant growth. I am glad we agree on the effect of CO2 in isolation. Now let's examine the combined effect. There are three things we need to consider: the direct effect of CO2, and the indirect effects of rising temperature and changing rainfall (I won't say rising or falling; more on that in a moment). Rising CO2, we agree, in isolation, stimulates plant growth. I think it's fair to say, in addition, that the effect is large. For a doubling of CO2, I would suggest that it would not be remotely credible to suggest that the stimulatory effect of CO2, in isolation, would average only 10%, across important crop species (however one defines that). I would suggest that even only 20% would be surprising, though it would be out of the range of "not remotely credible" at that point. Now, temperature. You and I both agree that a doubling of CO2 will produce a rise in global average temperature. The IPCC I believe estimates 3 degrees C. I think that's an overestimate but will accept it arguendo. Before making any prediction, simply an observation about current agriculture: the pattern is that the warmer the region, the greater the harvests tend to be, at least across most of the range of current temperatures on Earth. In terms of agricultural productivity, Nigeria > Mexico > France > Norway > Alaska > Greenland > Antarctica, for example. I am unaware of any evidence one way or the other about the hottest regions with reasonable amounts of water. It would be interesting to know that. I mean, obviously the Sahara has very little agriculture, but I think "dry" is the issue there more than "hot." So let's save that question for water. In light of this, I believe it is fair to say that the temperature increase, in isolation, is likely to produce an increase in agricultural productivity. It is possible that the hottest regions may suffer a loss; I am not sufficiently familiar with the evidence to answer that. But for the regions listed above, an increase seems certain for temperature zones from France on down and likely even at the Mexico level, ignoring water for the moment. Nigeria I don't know one way or the other. But overall, in light of this, it seems fair to expect that temperature would also produce an increase.

Now, changes in rainfall patterns. The IPCC talks from time to time about "more droughts." But have they made any effort to quantify whether or not we are currently seeing no droughts? I am unaware of any such effort. I find this curious; certainly the IPCC has shown that it can quantify a claim when it desires to. Furthermore, "more droughts" is a claim which could be quantified in terms of actual precipitation data. Make a mathematically reasonable definition of "drought" or "precipitation variability"; I don't care what it is. A yeare with less than 50% of the mean precipitation (drought); the standard deviation of precipitation divided by its mean level over a 20-year period ("precipitation variability"); whatever. Then, based on actual rainfall data, it would be possible, and I daresay not difficult, once one had gathered the data, to determine a trend. One could then determine what that trend is, and put it in the IPCC report.

Feel more than free to correct me on this, but as far as I am aware, the IPCC has not done any such thing. That suggests to me that the data do not show any unfavorable trend in droughts to date. Feel more than free to correct me on this if I am wrong. But if I am not wrong, then I will infer from that that so far, precipitation patterns have not become any worse than they were in the past. Because if they had, we sure as hell would have heard about it.

Therefore, overall, we have one change (CO2) which is uniformly favorable to harvests, and dramatically so. A second change (temp.) which is mostly favorable, or possibly entirely, depending on what happens at the hottest end of the scale. And a third one (rainfall) which I infer that data (as opposed to models) don't show overall worsening. Lastly, at high CO2, the plants need less water overall. So even some worsening in rainfall, if it were to happen, would be overwhelmed by that effect. Hence my "blindingly obvious". There is actually one more piece of evidence, weaker than the others IMO, it's possible there is another cause, though I'm not aware of one. That's the trend in the May-Oct. in the Hawaiian CO2 data. That drop has been increasing over time. Not uniformly, there are ups and downs, but the overall trend is clear. This drop is generally attributed to the Northern growing season. If the drop is increasing, the natural inference is that said growing season is getting stronger. It is possible there is another explanation, though I haven't heard it. CometEncke (talk) 18:04, 25 January 2016 (UTC)

How do you get from "an increase in CO2 from the base level to a moderately increased level alone increases primary plant productivity for most plants in situations where growths is not limited by the unavailability other resources" via "Rising CO2, we agree, in isolation, stimulates plant growth" (drops a lot of qualifiers) to "uniformly favorable to harvests, and dramatically so"? It's not just crop plants that profit - weeds do likewise. And "uniform" is the opposite of what the sources say - indeed, it's very much non-uniform. There is also very little data on plants grown in cultures (where they compete for resources). As for droughts: It's not just annual precipitation thats relevant, but precipitation at the right times of the year, not to mention the vanishing buffering capacity of vanishing glaciers. Winter snow that melts in spring is useless for irrigation in summer. As for the rest: I suggest you take a look at the IPCC reports instead of speculating about what they don't contain. WG2AR5C3 deals with hydrology, and WG1AR5C2 has information on the development of precipitation and the hydrological cycle. --Stephan Schulz (talk) 19:01, 25 January 2016 (UTC)
I did drop some qualifiers. It's a fair point. Sorry about that. The IPCC has lost my confidence with its repeated over-predictions. We are up to their 5th report now, and so far we are well below their temp. predictions for the first four, taking business-as-usual emissions, which is more or less what we've had. How many times can one cry wolf? Why go after *me* so hard; isn't this about finding truth? CometEncke (talk) 22:37, 25 January 2016 (UTC)
You might want to look at this page and the papers referred therein. Of course, the IPCC reports are current and evolving science, so their predictions are not perfect. But they are, contrary to cherry-picking contrarian claims, quite good. "This" thing we are doing is cutting through misinformation and misunderstanding. I'm a computer scientist - indeed, I'm an expert in a very small field of algorithm design, logic, search heuristics, with a smattering of knowledge about air traffic control and machine learning. I don't have the hubris to believe I can contribute significantly to finding "the truth", or even the increasingly better approximation to the truth that science gives us, in a field as wide and complex as climate science - at least not without overturning my career and starting again at an undergrad level. On the other hand, I have a decent layman's overview of the field, and I can sometimes recognise claims as plain wrong. If I see those, I try to correct them. I assume good faith, i.e. I assume that my debating partners will be happier to improve their understanding than to score debating points. On the other hand, if not, I have little sympathy. I can't stand e.g. creationists who serve the same over and over refuted claims over and over again. If you don't trust the IPCC, that's your prerogative. But making wrong claims about them and their reports is not. --Stephan Schulz (talk) 23:02, 25 January 2016 (UTC)
I'm sorry to see how this discussion has gone. My comments have been full of statements like "feel free to correct me on this, but . . .", or "I did do x. Sorry about that", or even "There might be another explanation, but I haven't seen one." (FWIW I did think of one possibility afterwards. No idea if it fits or not.) If the intent was to win me over, it's been done; it would have involved noticing more places where I was obviously correct, and making corrections gently, rather than focussing relentlessly on the negative, and above all, realizing that learning from the other guy is a two way street. So, in this case, you've lost me. You may feel that was a foregone conclusion. It wasn't, but it's a done deal now. Still, best wishes, and better luck next time. I'm a technical person too. I don't want to get specific, but you can be sure that if I believe something to be factually correct, I don't care in the slightest whether or not it fits with any opinion I have; I will revise my opinions, not facts. Random case in point, Adnan Syed from "Serial" is innocent. Not "not proven guilty." Proven not guilty, beyond a reasonable doubt. In time, the courts will agree, at least up to "not proven guilty." Bank on it. Kevin Urick should be in prison, but that will never happen. And best wishes, still. Really. CometEncke (talk) 00:42, 26 January 2016 (UTC)
Maybe we have different expectations. I'm not proselytising for a religion, trying to safe souls whatever the cost. Also see Richard Dawkins on a related topic of methodology. I had no idea who Adnan Syed is, nor why you brought him up, but I notice that you apparently have formed a very strong opinion on the issue - based on what? A polemic podcast? I'm not saying you are wrong, but I find no substantial support for you being right, either. --Stephan Schulz (talk) 15:13, 3 February 2016 (UTC)
I brought him up to make the point that I base opinions on facts and evidence, everything else be damned. My mind was actually pretty clear before the polemical (and dead right) cast came out. A friendly wager the courts agree and AS is out this year. CometEncke (talk) 22:07, 11 February 2016 (UTC)

Cold blob (North Atlantic)[edit]

Hi Stefan, i just created a new article, maybe you can have a look https://en.wikipedia.org/wiki/Cold_blob_(North_Atlantic) Thanks. prokaryotes (talk) 19:16, 22 January 2016 (UTC)

Sorry for the later answer - I'm busy with some real life science and questionable discussions ;-). I don't know enough about this to do a useful review off-the-cuff. If I find the time, I'll look later. --Stephan Schulz (talk) 22:08, 23 January 2016 (UTC)
Thanks, also new Climate action, and Climate change and national security. prokaryotes (talk) 18:16, 25 January 2016 (UTC)

Complexity of Addition in Finite Automata[edit]

I couldn't understand the finite automata shown in the Article on Finite Automata(P.65).It might be because I'm weak in binary mathematics.Addition is very simple.I can't get why the author has shown 'addition' as complex.Could you help me and give a brief explanation of each state given in the finite automata.JUSTIN JOHNS (talk) 06:49, 27 January 2016 (UTC)

@JUSTIN JOHNS: I'm rather busy at the moment, and the slides are missing the main part (the presenter). But as I understand it, the automaton is processing a sequence of 1-bit-additions (least significant bit first) and checks if the overall addition is correct (so it's not performing addition, but verifying addition). The alphabet of the automaton consists of the individual combinations of 3 bits (the first two are the input, the last is the lower order bit of the output), i.e. each of the characters of the alphabet is one of the 3-bit combinations (written as a vertical vector with a dash, but that is just syntactic sugar). As you probably know, a 1-bit addition has a two-bit result (1+1=2= (binary)10)). This extra bit is the "carry" bit, and if you build a multi-bit adder, you must take it as an additional input for the next bit slice - see full adder. The automaton starts in the state R0, and R0 says "the carry bit is zero". It then checks the character - there are 4 valid additions, and 4 invalid ones.The valid ones are 0+0=0, 1+0=1, 0+1=1, 1+1=10. The last one also sets the carry bit, which is why the automaton goes to the state R1 ("the carry bit is 1"). All other variants are wrong and lead to the error state. In R1, there are again four correct results, and 4 wrong ones. But since we now have the carry bit as an additional input, the correct results are 0+0+c=1, 0+1+c=10, 1+0+c=10, 1+1+c=11, and the other 4 are wrong. In the first case (0+0+c=1), the carry bit is consumed and we drop back to R0. And of course, once the addition is wrong, it stays wrong (which is why the error state goes to the error state with the whole alphabet). It's apparently "complex" to understand, which is why you have to ask ;-). On the other hand, it shows that verification of addition is at worst linear in the number of bits with respect to time complexity and constant with respect to space complexity. --Stephan Schulz (talk) 13:40, 27 January 2016 (UTC)

Yeah this answer gives me a sense of hope for my doubts.Could you list the four invalid ones for the state R0.JUSTIN JOHNS (talk) 06:55, 28 January 2016 (UTC)

Well, it's the combinations missing from the list I gave above: (00/1). (01/0), (10/0), (11/1). Now you list the wrong ones in R1! --Stephan Schulz (talk) 08:39, 28 January 2016 (UTC)

I think it would be 0+0+c=0,0+1+c=01,1+0+c=11,1+1+c=11 but I'm not sure.I really understood the wrong states you listed above by these two sentences you've mentioned:"the first two are the input, the last is the lower order bit of the output","so it's not performing addition, but verifying addition".Could you tell why the states in R1 are only 4 because we could also list 0+1+c=11 as a wrong state isn't it?JUSTIN JOHNS (talk) 10:01, 2 February 2016 (UTC)

I think we must first get the notation straight.The three states of the automaton are R1, R2, and error. The alphabet is a set of 8 letters, Σ = {00/0, 00/1, 01/0, 01/1, 10/0, 10/1, 11/0, 11/1}. Transitions take a state and a letter and produce a new state (which I will write (S,l ->S'). The transition table is
 Delta  | 00/0  | 01/0  | 10/0  | 11/0  | 00/1  | 01/1  | 10/1  | 11/1  |
 -------+-------+-------+-------+-------+-------+-------+-------+-------|
 R0     | R0    | error | error | R1    | error | R0    | R0    | error |
 R1     | error | R1    | R1    | error | R0    | error | error | R1    |
 error  | error | error | error | error | error | error | error | error |
If the carry is 1, i.e. if you are in R1, then 0+1+c=10 (decimal 2), so indeed (R1, 01/1 -> error). You got 1+1+c=11 wrong - if c is one, then 1+1+1=11 (decimal 3), i.e. (R1, 11/1 -> R1). --Stephan Schulz (talk) 13:33, 2 February 2016 (UTC)

Wow that's really great.Now I could understand the states present in the automata and their transitions.To be honest I really understood the answer after analyzing the document many times before you have posted the answer.Also I dont' have a reliable internet connection.So sorry for checking the answer too late.It's my mistake to ask a question without properly looking the document.Anyway that's a great help and thanks for your kindness and patience.JUSTIN JOHNS (talk) 07:47, 8 February 2016 (UTC)

No problem. You're welcome. --Stephan Schulz (talk) 10:42, 8 February 2016 (UTC)

Query[edit]

This seems kinda like it's in your bailiwick. Does it actually mean anything? Shock Brigade Harvester Boris (talk) 04:31, 26 February 2016 (UTC)

Well, for me it reads like Woowooowoooo, and Scientific Research Publishing does little to change my initial impression. The IF of the journal seems to be 0.00. --Stephan Schulz (talk) 11:43, 26 February 2016 (UTC)
Thanks. Yeah, I read it and thought "WTF?" And an article with over half the citations being to the author's own work usually isn't a good sign. But as I said, it's outside my field. (You may or may not want to comment here.) Shock Brigade Harvester Boris (talk) 14:24, 26 February 2016 (UTC)
I don't really think it's anyones field (not mine, to be sure), but rather something made up. There are some interesting analogies between entropy in physics and information theory, but this does not mean that you can translate every informal idea from some information domain into any arbitrary physics domain. I already commented at RS/N, maybe a bit to cynical... --Stephan Schulz (talk) 14:53, 26 February 2016 (UTC)

Environmental impact of the Wikimedia movement[edit]

Hi Stephan, I created an essay regarding the environmental impact of the Wikimedia movement on Meta and I am now looking for ideas regarding the project. I saw that you're interested in sustainability, so I'd love to hear your comments and maybe even have your support! Thanks, --Gnom (talk) 21:43, 24 March 2016 (UTC)

Pumping length of DFA[edit]

Do you mean to say that we can only use finite languages for pumping lemma?Could you tell the pumping length for this dfa?Does this DFA accept an infinite or finite language.Could you help me.JUSTIN JOHNS (talk) 08:27, 28 March 2016 (UTC)

DFA finite lang trie.png
Hi Justin. No, that's not what I'm saying. Maybe we should list a couple of facts to be sure there is no misunderstanding:
  1. An alphabet is a finite set of letters, and a word is a finite sequence of letters from an alphabet. In particular, all words have finite length.
  2. A language is a set of words. A language can be finite (as e.g L={a, ab, aab, abb, bab}) or infinite, as e.g. K={a, aa, aaa, aaaa, ...}. Note that all words in K are finite (indeed, words are finite by definition). However, in K there is no upper limit to the length of a word - for every bound, say n, there are words of greater length, e.g. ana (n times a, followed by one more a). This is true for all infinite languages.
  3. A language is regular if and only if there is a DFA that accepts it.
  4. All finite languages are regular. You can simply construct the DFA as a trie - see the one on the right for L. Please not that the DFA does not have a loop (i.e. there is no way to visit a state twice when processing one word).
  5. The pumping lemma says that all regular languages have a pumping length p, and that all words longer than p can be pumped. If such a word exists, then the language is necessarily infinite. All languages with pumpable words are infinite.
  6. But there is a second possibility, namely that there is a p so that there are no words of length greater than p. This is the case with all finite languages. The pumping lemma is true for finite languages because there is a p so that the condition on pumpable words becomes vacuous - you can pump all words of length greater than p, because there are no such words, i.e. you can pump all the 0 words of length greater than p.
As for your example: The automaton has a reachable loop that can lead into an accepting state (indeed, it has several such loops), hence it accepts an infinite language. It does, for example, accept all words of the form bia for any natural number i. I don't know a sharp boundary for p, but if you chose p as 4, you are on the safe side - with 4 transitions, the automaton has to visit at least one state twice, so any word of lengths 4 or larger has to go through a loop - and whenever you go through a loop once, you can go through it any number of times. I'm sure @BenRG: will step in and explain if I made any more errors ;-). --Stephan Schulz (talk) 12:26, 28 March 2016 (UTC)

That's a good description of pumping lemma and how we apply it to finite and infinite language.I could see that it's easy to find the pumping length p for finite languages since all we need to do is find a length greater than p such that there aren't any words so we can pump 0 words as you have mentioned.Could you tell is there any way to find the pumping length p for infinite languages.Do we only need to look if a state is revisited to find out it's pumping length p?JUSTIN JOHNS (talk) 07:03, 29 March 2016 (UTC)

Well, if you have a language and a DFA for it, the number of states gives you an upper bound for p. You can also minimise the DFA, possibly getting a better bound for p. I'm not aware of a universal method to find a sharp bound. But maybe you are on the wrong track altogether? The pumping lemma is most often not used to actually pump a language, but to show that a language is not regular, because it cannot be pumped on all long words. In that case, you assume that the language is regular, then you postulate the existence of p, and then construct a word in L that is longer than p, but cannot be pumped. For that, you don't use a concrete value like 3 or 5, but construct your word based on p. That contradicts the pumping lemma, and hence you know that the assumption that the language is regular has to be false. The canonical example is to use apbp to show that {aibi | i in N} is not regular (your would pump more copies of a into the word, thus breaking the symmetry). Note that apbp is much longer than p - n fact, it's twice as long. But also note that we don't use a fixed value for p (indeed, since the language is not regular, no such value exists). --Stephan Schulz (talk) 12:08, 29 March 2016 (UTC)

Thanks a lot for your effort.That made my doubt clear on how to find p for infinite languages.I was really stuck on whether p is actually the number of states in a DFA or is it any other number that takes you through a loop.Now after you told about the 'upper bound' it's sure that there won't be any p such that it's greater than the number of states.Could you tell why we enforce the condition |xy| ≤ m (number of states for dfa) for pumping lemma.In this article(slide 23) it tells that the condition is enforced because of unique states in 'xy'.I couldn't get the need of unique states in 'xy'.Could you help me.JUSTIN JOHNS (talk) 06:47, 30 March 2016 (UTC)

Well, you split the word into xyz. Conceptually, x is the part leading to the loop, y is what processed in the loop, and z is the trace from the exit of the loop to an accepting state. In the slides you linked to, m is essentially just another name for p. The argument goes as follows. L is regular. Therefore there is a DFA for L. That DFA has m states. Therefore any accepted word with m or more letters must visit a state twice, i.e. there is a loop in the automaton. We can use that loop multiple times to accept longer words. In the pumping lemma itself, we don't use the number of states, because we don't want to rely on a concrete automaton, but make an argument for every automaton that accepts L. --Stephan Schulz (talk) 13:46, 30 March 2016 (UTC)\

Okay.Could you tell why the restriction |xy| ≤ m can force y to have a special property as said in this lecture(p.68).JUSTIN JOHNS (talk) 07:26, 31 March 2016 (UTC)

Sorry - which slide and which property do you mean? --Stephan Schulz (talk) 07:33, 31 March 2016 (UTC)

Oh I mean to say slide 261 and it's better to read this lecture which says that the author tries to prove the language BALANCE(equal number of 0's and 1's) is regular using the Weak pumping lemma and he succeeds.While after he proves it using Strong pumping lemma I think he fails(not sure about this since I couldn't understand the proof).Then he points out that since the condition |xy| ≤ m is enforced it makes 'y' to have a certain property which I too couldn't understand.Could you tell why the author fails while using Strong pumping lemma.JUSTIN JOHNS (talk) 08:50, 31 March 2016 (UTC)

Ok. First, an important, indeed central, point. You cannot use the pumping lemma (weak or normal) to show that a language is regular. The pumping lemma states a sufficient, not a necessary condition on regular languages. All regular languages can be pumped, but there are also non-regular languages that can be pumped. Note that the Stanford argument is titled An Incorrect Proof and see slide 248. If a language is regular, sufficiently long words can be pumped. If this is false, the language is not regular. Now for your question: The weak pumping lemma is just based on the observation that if the word has more letters than the automaton states, processing the word must involve at least one loop, and the loop can be repeated. The "normal" pumping lemma for regular languages goes a little further - its based on the observation that you already need to complete at least one loop when processing the first p letters of the word. This extra constraint makes it easier (and sometimes possible) to construct the counterexample word. Check the proof for "balance" that follows, it's a nice example. --Stephan Schulz (talk) 09:27, 31 March 2016 (UTC)

Do you mean to say that using the normal pumping lemma we can prove that a language is not regular in a few steps?JUSTIN JOHNS (talk) 10:05, 31 March 2016 (UTC)

Either of the two (weak and "normal") pumping lemmas for regular languages can be used to show that some languages are not regular. Neither can show this for all non-regular languages. But the set of languages that can be proved non-regular with the weak version is a subset of those that can be proved non-regular with the normal pumping lemma. --Stephan Schulz (talk) 10:32, 31 March 2016 (UTC)

Even though I could get the difference between weak and normal pumping lemma I tried searching the web but couldn't understand the examples given over there.Could you give an simple example to show the difference between weak and normal pumping lemma.JUSTIN JOHNS (talk) 06:43, 1 April 2016 (UTC)

Check the "balance" example from your slides - all words that have the same number of a and b (I think its 0/1 in the slides). With the "weak" version, you don't know where to pump. So if you pick e.g. the word apbp, you could split it as x=ap-1, y=ab, z=bpz=bp-1, and you can indeed pump that (every time you get both an a and a b). But with the "normal" version, you know that x=ak, y=al, z=ambp with k+l <= p, k+l+m=p. Thus you would pump only copies of a into the word, breaking the balance. Hence balance is not regular. I've never seen the "weak" version before, so it's no wonder you have a hard time finding examples. --Stephan Schulz (talk) 08:27, 1 April 2016 (UTC)

Okay.Could you tell why we get "both an a and ab" in the weak version.I think it would only be 'ab' since y only contains 'ab'.JUSTIN JOHNS (talk) 08:54, 1 April 2016 (UTC)

There is a semantically important space. You get both "an a" and "a b", because "an" is the article that goes with "a" (which starts with a vowel sound), while "a" is the article that goes with a "b", which starts with a consonant.  ;-). --Stephan Schulz (talk) 09:26, 1 April 2016 (UTC)

Oh I see.I think you're trying to say that you've really mentioned "both an 'a' and a 'b'" instead of "both of 'a' and 'ab'".It's my fault of how I interpreted your sentence.Don't worry.I think the grammatical session is over.Could you tell whether the 'z' part is bp or is it bp-1?JUSTIN JOHNS (talk) 09:42, 1 April 2016 (UTC)

In the weak case, you know nearly nothing about z. In the non-weak part, you only know that the initial segment xy is less then p long, so (for the word apbp), both x and y consist just of as. The z part is the remaining as, which may be zero, and all the bs (I try to use "as" as the plural of a ;-). --Stephan Schulz (talk) 10:11, 1 April 2016 (UTC)

Sorry I would like to know if you have mistaken 'z' for bp instead of bp-1 so we get the expression apbp.If we use the 'z' part as bp I think we would get the expression as apbp+1 isn't it?JUSTIN JOHNS (talk) 10:23, 1 April 2016 (UTC)

Another typographical problem, I think. "l" is the lower case letter L, not the digit 1 in the above. --Stephan Schulz (talk) 10:51, 1 April 2016 (UTC)

Could you tell if we use "l" instead of "1" is it possible to get the expression apbp in the "weak" version.To be honest I couldn't see an "l" in "weak" version even though I could see "l" in "normal" version.JUSTIN JOHNS (talk) 06:45, 4 April 2016 (UTC)

Yes, sorry, I was confused about which part you referred to. Your conjecture was correct, I have corrected it above. --Stephan Schulz (talk) 07:17, 5 April 2016 (UTC)

Okay.Let's consider a finite automaton that has more than one loop(revisits the state entered again) like this dfa considered before.When we look this image we could see the possible loops are q0q1q2q0,q2q3q2 etc but this lecture(p.18) says that there's only one loop in q part.So could you tell if we could have more than one distinct loop in dfa as we have seen in the previous image.To check if a language is not regular do we need to look on all possible loops in the dfa to check whether it obey pumping lemma?JUSTIN JOHNS (talk) 06:33, 6 April 2016 (UTC)

No. The argument for the pumping lemma goes as follows: If a language is regular, then there is a DFA for it. We don't construct the DFA, we don't know the DFA, we just know that it exists and has a fixed finite number of states. Then, when the language has arbitrarily long words, for all "long enough" word, the automaton has to go into at least one loop, and that loop can be pumped. If there are more loops, the word can be pumped in different places, but that is nothing we need for the argument. --Stephan Schulz (talk) 09:30, 6 April 2016 (UTC)

Could you tell the steps we should take to show whether the language accepted by this automaton is not regular.While looking this image I can see loop in many places so don't know which all states we have to make x,y and z to check whether the y part contains a loop.JUSTIN JOHNS (talk) 10:32, 6 April 2016 (UTC)

Sure. The language accepted by the automaton is accepted by a DFA (namely the very one ;-). Hence it is regular by definition (or by equivalency of regular expressions and DFAs, depending on how you define regular languages). So I would take no steps at all to show that it is not regular, as that would be an exercise in futility. --Stephan Schulz (talk) 12:18, 6 April 2016 (UTC)

Oh sorry but I still can't get how to know whether an automaton is a DFA.Do we only need to check if an automaton accepts an regular expression to check whether it's an DFA?If that's true should we rely on Pumping lemma to check whether an automaton doesn't accept an regular language and in turn proving that the automaton is not a DFA?JUSTIN JOHNS (talk) 06:43, 7 April 2016 (UTC)

The definition is at Deterministic finite automaton. Basically, if you only have finitely many states, no other memory, and the transition relation is a function (i.e. for every state/letter combination there is a unique successor state), you have a DFA. If only the first two are true, you have an NFA, but NFAs can be converted to equivalent DFAs, so both automata classes accept exactly the regular languages. Other automata in this field are push-down automata (which have an extra stack and can recognise context-free languages) and Turing machines (which have a read/write tape and can accept recursively enumerable languages). Also see Chomsky hierarchy. --Stephan Schulz (talk) 07:38, 7 April 2016 (UTC)

Yeah I think I was really confused between languages and DFA.I thought that pumping lemma is used to check whether an automaton is a DFA.Now only I realized it is used to find whether a language is not regular.When we use the pumping lemma can we use any DFA to check whether a language is not regular.JUSTIN JOHNS (talk) 08:10, 7 April 2016 (UTC)

And you are still confused, if less so. The pumping lemma states a necessary, not a sufficient condition for regularity. So if a language cannot be pumped, its not regular. But if it can be pumped, we cannot say anything, as there are non-regular languages that can also be pumped. And the "any DFA" is at least ambiguous. What we do (when we go back to the underlying argument) is to a assume that a DFA exists and then show that this leads to a contradiction. Hence there is no DFA, hence the language is not regular. Having the pumping lemme in an abstract from saves us from arguing on that level every time. --Stephan Schulz (talk) 12:10, 7 April 2016 (UTC)

I think altogether pumping lemma is used to check whether a language is not regular and that sum of states in x and y shouldn't be greater than the pumping length isn't it?I really couldn't get why the condition |xy| ≤ m is enforced for pumping length 'm/p'.If the condition |xy| ≤ m is violated would it be an issue in pumping lemma?Could you help me.JUSTIN JOHNS (talk) 06:37, 11 April 2016 (UTC)

I'm not really happy with the word "check", as that can be read to imply that you get a guaranteed result. With the pumping lemma we only get "not regular" or "don't know". As for the rest: the stricter the conditions, the easier it is to show that a language violates them, so the more powerful the lemma. We can impose |xy| ≤ m, because xy is the initial part of a word that is longer than than the pumping length, and hence already has to have at least one loop. --Stephan Schulz (talk) 07:10, 11 April 2016 (UTC)

Could you tell whether your last sentence violated the pumping lemma.I think you have to say that it's "no longer than the pumping length".Perhaps that's not a problem.Now only I could get that we impose the condition to speed the pumping lemma method.If we haven't imposed the condition could we still state that a language is not regular in less speed compared to the normal pumping lemma?JUSTIN JOHNS (talk) 08:15, 11 April 2016 (UTC)

I think my sentence is correct as stated. I don't know what you mean by "speed", but maybe it helps if you actually use the PL to show a couple of languages to be non-regular. The canonical example is, of course {aibi| i in N}, but you can also try {aibwbai | i in N, w an arbitrary word over Sigma}.--Stephan Schulz (talk) 08:34, 11 April 2016 (UTC)

I meant to say speed as " more faster the less it takes to prove a language is non-regular using pumping lemma".Could you tell whether the initial part of the word 'xy' should be longer than the pumping length.JUSTIN JOHNS (talk) 08:43, 11 April 2016 (UTC)

It's not a value judgement, so "should" is not quite appropriate. But you can pick a word such that you can pump it within the first p characters, where p is the pumping length. Take a look at the pumping game for some experience and a feeling for where there is a universal and where there is an existential quantifier. --Stephan Schulz (talk) 09:22, 11 April 2016 (UTC)

Sorry but I'm trying to ask whether the pumping length m should be greater than the sum of states in x and y or whether the pumping length m should be lesser than the sum of states in x and y in accordance with pumping lemma.JUSTIN JOHNS (talk) 09:35, 11 April 2016 (UTC)

It's the other way round! You propose that the language is regular. In this case, there is a DFA accepting it (by definition). That has a certain number (say p) of states. So any word of length p or greater has to run though at least one non-empty loop when processing the first p characters. The initial p characters of the word make up your xy, and the "non-empty loop" is the y on its own. Then you can repeat (or leave off) the y. If that makes you leave the language, then your assumption was wrong, and the language is not regular. You don't really need to think about DFA's at all when applying the pumping lemma, just when proving or explaining it. And in particular, you don't need to think about particular DFA's. --Stephan Schulz (talk) 09:53, 11 April 2016 (UTC)

Yeah maybe now I'm on the road.So we're just analyzing the first p characters of a word.Till know I've related 'x' to the initial and states till 'y' and 'y' to the _non empty loop_ and 'z' to the rest of the states left.I think this type of comparison would put you in danger isn't it?It's better to think based on the position of the word.Now we have the word should we look whether the initial p characters of the word repeats?I'm done with the word but still don't know how to analyze the word using pumping lemma to prove a language is not regular.JUSTIN JOHNS (talk) 10:38, 11 April 2016 (UTC)

Have you tried the pumping game I linked above? No, it's not just the first p characters that repeat, but some subsequence. In the game view, your opponent proposes a p, then you pick a word (usually constructed using p), then you opponent gets to split it, and then you can repeat the y part as often as you like. If you can manage to create a word that is not in L, you win (the language is not regular). Usually, the hard part is coming up with a good word, most of the rest is easy. --Stephan Schulz (talk) 23:35, 16 April 2016 (UTC)

Yes of course I tried it.Hardly to say if there's any language that needs memory it can't be a regular language and to prove this we use pumping lemma isn't it?Is there any language that doesn't need memory but is not in the class of regular languages?If there isn't anyone then all the languages that don't need memory would be in the class of regular languages isn't it? JUSTIN JOHNS (talk) 06:25, 18 April 2016 (UTC)

Deepness in the Sky = Exploitation[edit]

If this is about overlong plot summaries or something, then sure, we can stop them from being overlong. But otherwise, A Deepness in the Sky meets all possible criteria for being exploitation fiction. It's a pedantic and unconventional thing to say, but it's true. It's good exploitation, though, I didn't mean to say the novel is bad exploitation. Bad exploitation doesn't win Hugo Awards. Maybe I should watch Sweet Sweetback's Badass... Song (Q1812665). If I fall in love with the genre, then I should read award-winning hard SF. Did you revert my edit under the false premise that it was unadmiring? 203.215.119.40 (talk) 12:18, 3 May 2016 (UTC)

Not at all. I just don't see how "Deepness" meets the criteria for the exploitation genre. It does not "exploit sex, violence, drugs, or other elements meant to attract readers primarily by arousing prurient interest". Sex, drugs and violence to occur, but they don't dominate, and they don't "arouse prurient interest". Otherwise, you could label War and Peace as exploitation fiction, too, or even the Bible. If you want to insist on the point, bring a reliable source making the connection directly. --Stephan Schulz (talk) 15:20, 3 May 2016 (UTC)
I'll assume you meant "Sex, drugs and violence do occur". The only thing you've convinced me of is that my contribution is original research. And that's all you need to convince me of. I've actually read A Deepness in the Sky. We can discuss this, but I don't really see the point, because no amount of discussion would change the article. I'm sorry I forgot about the original research rule. 110.55.0.4 (talk) 06:40, 4 May 2016 (UTC)
Thanks. I'm always sorry to resist well-intend contributions, but in this case I think I'm right. I'm glad you accept the NOR rule. Reasonable people can disagree, and personal competence is hard to ascertain on Wikipedia - which is why we ask for reliable external sources. --Stephan Schulz (talk) 09:02, 4 May 2016 (UTC)
There's another problem with the article. Editorializing. There's glittering generalities everywhere, and we already know the Emergents are the bad guys, we don't have to be told that. Can we get rid of the word "totalitarianism", or is WP:EDITORIALIZING one of those rules we don't follow? 110.55.1.110 (talk) 10:59, 17 May 2016 (UTC)

Request for Access in Libertarian Party presidential primaries, 2016[edit]

Hello! You recently did a full protection on the Libertarian Party presidential primaries, 2016 Wikipedia page. I came to the page to update one of the polls, but as it turns out, someone removed it for being "inaccurate". I would assume good faith, but as the only poll removed by this person was the most recent one showing Johnson trailing Petersen it seems to have been vandalism. I would request that the removed polls be restored to the page. However, since there were two polls removed (and I'm not sure who removed the other one), a copy of all the polling data can be found on Gary Johnson presidential campaign, 2016. Thanks, and have a good day. SirLagsalott (talk) 15:17, 8 May 2016 (UTC)

Hi SirLagsalott, please discuss any change on talk: Libertarian Party presidential primaries, 2016. If there is consensus for the edit, use Template:Edit fully-protected to request the edit. Or wait till tomorrow - I only protected for 24 hours to take the momentum out of the edit war. --Stephan Schulz (talk) 16:04, 8 May 2016 (UTC)

Collect essay; second bite at the cherry[edit]

You participated in an MfD discussion about an essay by Collect that was in mainspace. The result was userfy and it was moved to user space accordingly. The essay has been moved back to mainspace. There is a discussion as to whether it should be renamed and moved. The discussion is here. Writegeist (talk) 00:38, 17 May 2016 (UTC)

Precious anniversary[edit]

A year ago ...
Cornflower blue Yogo sapphire.jpg
"I found the instant
improvability of Wikipedia
to be nearly irresistible"
... you were recipient
no. 1125 of Precious,
a prize of QAI!

--Gerda Arendt (talk) 13:54, 31 May 2016 (UTC)

Very good! Thanks for the reminder! --Stephan Schulz (talk) 14:13, 31 May 2016 (UTC)

Buch[edit]

Thanks for that :-). HD is a rather regrettable case; I'd say more except there are some slight parallels to mine :-( William M. Connolley (talk) 09:09, 28 June 2016 (UTC)

I had the impression that this was not purely motivated by what's best for the encyclopaedia. And I know that I've got more "patience with questionable characters" barn stars than you ;-). --Stephan Schulz (talk) 09:35, 28 June 2016 (UTC)
There won't be more problems like that for a while[3]. Shock Brigade Harvester Boris (talk) 12:36, 28 June 2016 (UTC)
Oh, go on, fix the header. I would, but it wouldn't be welcome William M. Connolley (talk) 14:17, 28 June 2016 (UTC)
I'm confused - which header? I'm also busy: http://www.cs.miami.edu/~tptp/CASC/J8/ (running the E theorem prover ;-). --Stephan Schulz (talk) 09:34, 29 June 2016 (UTC)
You mean the speedy? That was already taken care of. Shock Brigade Harvester Boris (talk) 13:04, 29 June 2016 (UTC)
No, I meant [4] which I've now done; sorry if I was obscure William M. Connolley (talk) 13:47, 29 June 2016 (UTC)
HD has reacted with that good humour and common sense so characteristic of him [5] William M. Connolley (talk) 14:53, 29 June 2016 (UTC)
At least he is terse... --Stephan Schulz (talk) 09:49, 30 June 2016 (UTC)

Islam El Shehaby and Or Sasson[edit]

Hello, Stephan Schulz. I am contacting you to let you know that there are some problems cropping up on the Or Sasson and Islam El Shehaby articles. We have at least one user, User:Or Sasson, who claims to be at least related to one of the subjects of the articles and at least fifteen reverts between both articles, including one by myself. I know that you are at least active on the Islam El Shehaby article, but I am less certain if you are involved with the Or Sasson article. Additionally, it seems like the topic is becoming a larger issue that might need more eyes to look into and attempt to come up with resolutions. --Super Goku V (talk) 00:12, 14 August 2016 (UTC)

I only stumbled over both articles by accident, and try to keep the situation from overboiling. It seems to have cooled down (maybe only because of time zones and nighttime, but we'll see). I've left a message at User talk: Or Sasson. 11:30, 14 August 2016 (UTC)
Actually, it looks as if things are cooling down overall, but thank you for helping to keep things in order. --Super Goku V (talk) 22:25, 16 August 2016 (UTC)

Animal Farm in German[edit]

Hi! Are you interested in Farm der Tiere? Somebody's upset about alleged bias. YoPienso (talk) 07:07, 21 August 2016 (UTC)

Sorry, I'm offline for the next week - cycling without electronics, and don't have time to become involved now. If it's still relevant in 10 days, and I don't turn up on my own. --Stephan Schulz (talk) 19:57, 21 August 2016 (UTC)