Talk:LR(*) parser

Delete

Please delete this page. Thank you. — Preceding unsigned comment added by Paulbmann (talk • contribs) 01:09, 1 April 2018 (UTC)

@Paulbmann: We might delete it, but you on the other hand, can not. So, here is the thing, I nominated this page for deletion in accordance with our policies. If the admins would see it not worthy of inclusion we will delete it. My suspicion is that we might just merge it into another LR article so that this and that would be covered.--Biografer (talk) 02:03, 1 April 2018 (UTC)

Merge proposition

Considering that the article is notable but is unsourced, it talks about the same thing as LR parser with minor differences (such as an odd (*)). Any suggestions will be welcomed.--Biografer (talk) 03:40, 1 April 2018 (UTC)

It is hard to say for sure without any sources, but the ANTLR parser uses an LL(*) strategy, where the * represents a variable lookahead[1]. This article seems to be proposing something similar for an LR parser. There was a product LRSTAR, now defunct, that did something like an LR(*) parse. But without sources, and I haven't yet found any, I don't think we can merge. --Mark viking (talk) 05:15, 1 April 2018 (UTC)

@Mark viking: Great. So if there are no sources, then why can't just we simply delete the thing for now, and maybe somebody will recreate it in the future?--Biografer (talk) 05:37, 1 April 2018 (UTC)

You see, here is the thing: The initial creation of the article was by user @Paulbmann: who immediately (after apparent realization that he probably duplicated an article) deleted it. I reverted it, explaining in the edit summary that he have no authority to do it (at first, assuming that he was deleting a random article). Upon realizing that an article is virtually unsourced (and I too was unable to find any), I nominated it for deletion, explaining that the article by itself is not notable (for an obvious reasons, or at least I thought they were obvious), only to be greeted by this on my talkpage by user @Oiyarbepsy:. So, as it stands, we cannot merge it (because there are no sources to back up the facts, as you said), we cannot delete it, because one editor here doesn't like the idea, and we can't keep unsourced content, (that is in addition, is an original research (hence no sources) and promotional) because we have a policy now (as of 2010) that our articles must have at least one source in order for them not to be deleted. So, what shall we do? It can't be kept unsourced for long, can it?--Biografer (talk) 06:10, 1 April 2018 (UTC)

• To set some things straight, I had actually PRODed this as original research because Paulbmann posted on this talk page (comment currently erased but in history) that he invented this algorithm himself - but then posted on my talk page stating that it had, in fact, been in use for decades, including in government. So, part of the question, and why I announced this at Wikiproject computers, is whether this is original research or not. If this does, in fact, turn out to be original research, I am willing to restore the PROD, but only if a redirect does not make sense. But original research and not notable are not ever reasons for speedy delete, ever. Oiyarbepsy (talk) 16:07, 1 April 2018 (UTC)
  • Thanks for the background. I think everyone here is acting in good faith, so it just remains to come to a consensus about what to do. Digging further into LRSTAR [2], it is actually LALR. It tries LALR(1) and expands to LALR(K) if needed. Unfortunately, beyond blog posts and mailing lists, I haven't found any independent RS for LRSTAR as an exemplar of this method. Nonetheless, redirects are cheap and I'd support a redirect to LALR parser based on whats out there. A merge without sources seems like a bad idea, but converting to a redirect would still preserve the content article history if sources became available. --Mark viking (talk) 18:48, 1 April 2018 (UTC)

Mostly off-topic post moved elsewhere Oiyarbepsy (talk) 02:21, 2 April 2018 (UTC)

• • Mark vikingBiografer, do we have agreement that a redirect to LALR parser is a good solution? Oiyarbepsy (talk) 02:21, 2 April 2018 (UTC)
    • No objection from my side.--Biografer (talk) 02:24, 2 April 2018 (UTC)
    • I agree this is a good solution. —Mark viking (talk) 03:20, 2 April 2018 (UTC)