New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Finalise wiki link format #337
Comments
hmmm, |
It's not just the work of typing that we should be concerned with, it's also the cognitive overhead of links appearing unexpectedly versus explicitly forcing a link. The trouble with complex rules is that the user has to be able to rerun the rules in their head in order to understand what's going on. |
I think the rules are not complex. Ther's just a bug with the "word" detection
So for me this is just a word detection bug The question is, if "underline" and "hyphens" combine to words, so the parser sees them as one. In german language there are rules, that are not "optional" (rule 26-30)[1] ... If we apply the first assumption hypens and underline combine words we get perfect CamelCase words. Those words can be escaped by The problem I see, is that adding I'd use IMO it depends on the usecase, if Mund-zu-Mund-Beatmung is a CamelCase word or not.
So I really think it depends very much on the usecase. [1] http://www.duden.de/sprachwissen/rechtschreibregeln/bindestrich |
Hi... I hope you won't mind a little input from a very appreciative and fairly new user. Regarding this:
In the interest of simplicity (and interoperability), though, it's better to have fewer settings to toggle, so I'd actually vote for no inference at all. But of course you have the question of backward compatibility to consider. I believe that 'no inference' (i.e. explicit links only) is the Markdown (and the MediaWiki) way, and I am very keen (for reasons of interoperability etc.) on seeing increased support for that format in TW5; I actually visited here today to raise what turns out to be #352! Reasons for finding link inference annoying:
I hope these viewpoints are of some use, though I'm sure they are obvious to you anyway. |
Hi @pipedelimited that's very useful feedback, thank you. I think you might like the proposal in #345. It's basically a way to allow authors to control the wikification of their own tiddlers whilst giving them interoperability so that their content can be displayed properly in other wikis. |
Just a quick word from another somewhat new user to say that I totally agree with @pipedelimited - from my personal experience with wikis, I always end up disabling CamelCase links. Emphasis is already expressed by enclosing text between double chars - isn't it logical to do the same for links ? |
+1 for CamelCase and ~tildes, very annoying when using lots of them |
IMO this is a bug and needs to be fixed anyway |
A related discussion at the group: https://groups.google.com/forum/?fromgroups=#!topic/tiddlywiki/jE46tM1RzPQ some infos about the unicode chars: http://www.fileformat.info/info/unicode/block/latin_supplement/images.htm |
For the record, if one of my TWClassic wikis isn't running DisableWikiLinksPlugin from TiddlyTools, it's either because I forgot or because I haven't touched it since I discovered that plugin. |
I would like to provide this example image (sentences from German Wikipedia article about Wikipedia itself) to show that this is a real issue in our language. Here are six undesired links showing up in four sentences: I vote for no inference at all, too. Another reason is that practically none of the links do automatically match on Wiki entries, even if they exist, due to lingual transformations (singular, plural, genitive, compound nouns, …) and it seems to me to be the common case that you need the “pipe”, like |
I'll try to include this ticket in 5.0.14. My view is that automatic wikilinking should only occur for classical camelcase words. (In your example above @matthias-ronge, "ShareAlike" would be the only wikified word). |
@Jermolene |
“HelloThere” in “My-HelloThere” shouldn’t be wikified. Part of #337
I've made a series of changes to the camelcase rules for 5.0.14:
The idea is that these changes take us back to a much stricter, more conservative recognition scheme, with fewer false positives. |
Sounds nice, when can we expect 5.0.14 to be out? |
Thanks @mindfaQ - 5.0.14 will be out in the next couple of days. |
I did a short test and it seems nice till now :) thx a lot! |
Great, glad to hear that @pmario |
Just a quick word on disabling wikilinks - for a project I have to generate tiddlers automatically from an external source which is not wiki-aware, and I get a few false wikilinks. Not many, but enough to make my sense of consistency tingle. As regular imports need to be done, it is not feasible to correct the link manually. Correcting them automatically would mean to reimplement tiddlywiki's wikilink rules in my generator, which is boring. |
Hi @tgirod Your importer can add the following line to the imported tiddler.
How do your false positive links look like? |
Hi @pmario A mix of uppercase letters and numbers can appear as a designation code, or a password, or something else and should not be wikified by default, I think. |
I think, you should create a new issue for this one. |
TiddlyWiki5 currently uses the same regular expression for matching WikiLinks as TiddlyWiki Classic:
In plain language:
There are several obvious problems:
Generally, though, the rules are arguably far too loose. I think it makes sense for the wikilink rules to be on the conservative side, as it is easier to explicitly link a text than it is to suppress a wiki link.
The text was updated successfully, but these errors were encountered: