Re: [cowiki-dev] [cowiki] Using AJAX for form validation
Archie Campbell <[email protected]> Wed, 22 Jun 2005 15:53:45 +0100
| Newsgroups | gmane.comp.php.cowiki.devel |
|---|---|
| Message-ID | <[email protected]> |
For a purer sense of flexibility, we should allow bad wiki, in case there was no other way to express ones self. Otherwise we invite cumulative load-save errors-of-circumstance where a preview would have alerted all but the most illiterate user. For example, a table is left open, like so... <table> |- | I'm behaving well but | I'm going to forget to close this table ++ And on with the show... ...and how do we decide where to place the </table> closing element? At which point (during a parse) does the corrective code kick in? Do we walk the xml tree from the end to the beginning, looking-"ahead" for the best place to make an assumption? Assumption is the mother of all fuckups; let's leave correction to eyeballing a preview - the customer is always right. Regards, Archie PAUL HANCHETT wrote: > I think I understand what you're saying Archie, but I don't know that > I agree. There are quite a few elements that are required to be > paired, if they aren't the text won't be interpreted at all. Now, I > suppose that we could add the missing elements to "correct" the markup... > > Archie Campbell wrote: > >> The thing with wiki markup is that the parser is required to accept >> just about anything, whereas syntax-checking requires the entire >> document to be grammatically defined. That is, the parser can't >> declare an error at such-and-such a place without expecting a given >> syntactic element at that place, which would restrict the set of >> parsable wiki documents to those having a set grammatic structure. >> >> Bottom line... it can't be done. If it could be done, it would cause >> all coWiki documents to be boring. >> >> Preview, on the other hand, would be fairly simple and wholly useful. >> >> Regards, >> >> Archie >> >> PAUL HANCHETT wrote: >> >>> Using AJAX to validate the page seems like a reasonable use, >>> especially if we can also use it to identify where the trouble spot >>> is. I think it would be a good thing because then we wouldn't have >>> to save syntactically incorrect pages. >>> >>> I like the idea of offering a menu of valid links to insert. Such a >>> thing could also be used to validate links once they are entered... >>> >>> Using it for page maintenance also seems to make sense... >>> >>> Should we think about including a WYSIWYG editor as well? Clearly >>> it can be done... >>> >>> Daniel T. Gorski wrote: >>> >>>>On 18 Jun 14:06, John (yt) Hogenmiller wrote: >>>> >>>> >>>> >>>>>On 6/18/05, Sy <[email protected]> wrote: >>>>> >>>>> >>>> >>>> >>>> >>>>>>At the very least, AJAX lets bandwidth usage be much more intelligent. >>>>>>Sortof like what frames was supposed to do. ;) >>>>>> >>>>>> >>>>>> >>>>>Technically, AJAX (and methods like it) submit/retreive something >>>>>every time you exit a form element. But it's a very small amount >>>>>being submitted that typically, you barely notice. >>>>> >>>>> >>>> >>>>At first I need to say that coWiki didn't require JavaScript so far. If you >>>>want to change this basically, make it good. Means: if you want to use JS >>>>someday and take it for granted for coWiki to work, you may replace some >>>>functionalities with JS directly. These may be e.g. already existing >>>>dialogues. >>>> >>>>Such (XML)-RPCs/requests like this AJAX (horrible name for an old >>>>technology) provides, make sense in at least two coWiki cases: >>>> >>>> a) The display of the tree if you want to move documents. To select the >>>> target branch you may use XML-RPC to retrieve the subtree data to be >>>> displayed. This would also scale in the case that the tree has > 10.000 >>>> nodes or such. >>>> >>>> b) If you have a smart textarea that displays an additionaly layer with >>>> document name suggestions after you typed "((" or "[[" and a few >>>> characters. I think I don't need to describe this functionality further. >>>> >>>>regards dtg >>>> >>>>--------------------------------------------------------------------- >>>>To unsubscribe, e-mail: [email protected] >>>>For additional commands, e-mail: [email protected] >>>> >>>> >>>>