Clemens, It looks like you mixup 1) and 2), did you ?
On Thu, Jan 23, 2014 at 5:21 PM, Clemens Klein-Robbenhaar < [email protected]> wrote: > On 01/23/2014 02:49 PM, Sergiu Dumitriu wrote: > > On 01/23/2014 06:11 AM, [email protected] wrote: > >> Hi devs, > >> > >> I’m working to fix http://jira.xwiki.org/browse/XWIKI-9910 but before > I can fix it we need to decide something since we have 2 possibilities. > >> > >> - Option 1: The hidden flag is set at document translation level which > means when the user check the hidden flag it’s only for the current > translation > >> - Option 2: The hidden flag is set at the default document level (not > set at translated doc level) which means there’s a single hidden flag > >> > >> ATM the problem with XWIKI-9910 is that when the user checks the > hidden flag, it’s set at the translation level but when a translation is > displayed the value shown is the one from the default document. > >> > >> Option 1 offers more use cases but: > >> - users may be surprised > >> - users need to be careful to edit the default doc if they wish to set > the doc as hidden for all translations > >> > >> I’m not sure what option I prefer. Initially I was more for option 2 > but I’m now hesitating and leaning more towards option 1. Note that option > 2 means one more DB upate when saving a translated doc. > > > > I'm not sure 2 is going to work that easily, since by default queries > > don't filter by the "translation" flag. 2 means that we have to change > > every query (impossible if we count user queries), or the way the search > > APIs work (backwards incompatible). > > > > So +1 for 1. > > > > When first reading the original proposal I was more for 2) but I have not > thought about the queries. > Then I thought more about the queries and I feel 2) might still be better, > even though it breaks backwards comaptibility. > But then again I wondered about the relase that wants to get out of the > door, and feel that 1) is betterfor noe > and 2) might be added as a "new feature" later on. > > > about 2): Looking at the database structure might such a change make the > queries actually simpler and faster? > I.e. "give me all non-hidden docs in language X" : now it needs to fetch > the default document from the xwikidoc > beside of the language variant to access the hidden field, but with Option > 2 it needs only the "current language" > > About Backwards compatibility: > I guess users who are smart enough to be able to wrote HQL are hopefully > able to read release notes > and update their queries (or to accept that such stuff breaks). > > The bigger risk is that all hidden flags in all translations of hidden > documents need to be updated everywhere, > and the queries need to be updated. This is actually looking like 2) might > introduce a trail of little bugs ... > normally that should be ok, but maybe not the best idea for wanting to > have a stable 5.4 release soon? > > > From the users point of view I feel we have less confusion with 1) even > though it is less flexible. > Normal users never need to define "hidden" documents, because it is only > meant for "technical things", > but they might check that box just to figure out what it does anyway, > forget about it and create confusion later ;) > So the less complex that thing is, the better, and having only one > "hidden" flag makes it easier to figure > why that document X does not longer show up in the search ... > > > So unless I greatly overestimate the issues with rewriting the queries I > am more for 1) > > > > Use case: the master document is visible, and it is an important one > > (legal contract, license, official documentation...). Translations are > > being worked on. While a translation isn't approved, they'd like it to > > be hidden. > > > > UX proposal: > > > > - when a translation is created, it copies the hidden field from the > master > > - when a user changes the master's hidden status, a dialog shows up > > asking if all the translations should be changed as well or not > > - when a user changes a translation's hidden status, a dialog shows up > > asking for a confirmation if it's different from the master, warning > > about the possible issues caused by a difference in the flag > > - we display the hidden status of the translation in the UI > > > > On the other hand the UX for case 1) is simpler: > - if editing the default version of the document, keep current UI > - if editing a translation, show a text displaying the value of the > hidden flag > and a note that this can be changed in the default language. > > BTW: Would it be much work to hide the "Hidden flag" UI from "simple" > users, btw? > (If this affects the "save" method because currently not submitting the > value > for the hide-flag makes the save method assuming that is not set, then > just forget about it now.) > > > Clemens > _______________________________________________ > devs mailing list > [email protected] > http://lists.xwiki.org/mailman/listinfo/devs > -- Denis Gervalle SOFTEC sa - CEO _______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs
