As we are approaching the end of GSoC, little surprises along the way… more of this in the next entry, the next week, when the deadline will be passed.
16/08/2009 by Siv
As we are approaching the end of GSoC, little surprises along the way… more of this in the next entry, the next week, when the deadline will be passed.
That looks nice. Is this dialog only for KOffice or is for KDE in general?
I’m running KOffice 2.0.2 on Gentoo and there is no such dialog. Is this some internal patched version?
I don’t think that there should be a field for the mime-type imo that should be automatically set depending on the file-type.
Other than that imo the dialog seems too clouded. Maybe adding a seperator here and there could make it more clear.
Well, it will be for KDE in general, but for now it is a _very_ internal (that is: it exists only in my computer 🙂 ) patch/hack to get that; I may write more extensively the next week.
About the UI: yes, there will probably be some modifications, also from a usability point of view, after the end of GSoC; about the mimetype, remember that you always have a mimetype automatically set, also in the standard dialog, in the format drop-down list; maybe I gave it a not so clear name…
Thanks for explaining. After browsing some more through your blog I kind of figured it out as well, but wasn’t sure since I’ve been running 2.0.2 a few days before it was officially announced (yes, Gentoo is up to date again 😉 hehe).
As far as MIME types are concerned, I still maintain a view that KDE could do a lot better then it does now. It seems like the current implementation only (or at least mainly) takes into consideration the file extention. While even if you try a small CLI tool like ‘file’ — which should be present on pretty much every GNU/Linux system — can guess MIME types by looking at their contents and do so a lot more accurately.
Hey Sivvy,
At first I was a bit sceptical about this whole approach – theoretically it makes sense in a modern world but it just didn’t look right. Now I must admit I’m impressed, you’ve won me over 😉
Great work, please finish it and get it in KDE some day 😀
that really looks cool – thumbs up!
I admire the effort and thought gone into this, well done, but…
I kind of disappointed we still have file-open, file-save, file-save-as this was a concept to help computers (of the 1980s) with not enough floppy disk drives. What are you trying to do – make a copy of the current document with a new name?
The Name: should be filled in, even if it is only a name based on the title, top metadata item, current date and time. Spell check or dictionary services should be available for all input fields.
I think Metadata should be renamed tags (not that it makes anymore sense). The Metadata box should be filled in this the suggested and recent tags/metadata and user can vote them up or down if they feel inclined. The Suggestions and Recently used boxes can be removed. I think it would make more sense to have the meta data/tags as a side panel tool box, the prototype is moving away from a document metaphor towards ramming the file-system down the (non-technical) users throat.
Type, Mime type, and description are kind of all the same thing. It should default to KWord’s default format (or current format assuming this is copy with name change) and perhaps allow you to export (read only) copies to other word-processors. Thinking about it, it should just save in its feature complete native format, but there is a service running that can translate to other formats, like when you put a audio CD in the CD drive, and it appears as folders of MP3s, Oggs, WAVs, etc.
@Robert: well, I think the open/save/save-as problem that you name is somewhat more general, the dialog you see is just a parallel to the current existing system; it may be interesting to change this paradigm, but we should think about the possible substitution…
The metadata, IMHO, cannot be renamed tags, because it does not only contain tags, it may contain the current context, or affiliated projects/documents… it could be interesting, anyway, to have a single view with voting.
Mime type and type are not the same thing: meta type is the legacy system used to choose the document type when saving, and I think will be used for a long time, the type is the real type of a document, like a generic “Document”, or you can create a subclass like “Holidays” and associate to that type all the docs (but also pictures, videos…) related to your holidays…
Your last paragraph about the ‘mime type’ is really interesting – a sort of keywords and categories panel/dialogue might be useful to the non-technical user. I think it would need to be semi-automatic – the description filled in with a template description and the user selects the appropriate words. I think icons for the categories/keywords would speed up human recognition, perhaps a selection of Pinheiro’s lovely icons could be provided.
Of-course, different users will have different needs, and some may not be able to adjust to not having Save and Save As on the file menu. Ideally, my computer could suffer a power cut, and I could return to my document without the torturous dialogue of OpenOffice.org, Google Docs seems to save as you go along – you can close your browser without fuss.
I wasn’t entirely happy with Tags (for Meta Data), I (think I) know what meta data is, but many non-technical users seemed confused by the term. Perhaps these applications just need a translation into non-technical English dialect. What is meta data from a user’s point of view, it is like notes scribbled on the folder/cover of the document. Where does the computer store it – is it like META elements in the HEAD of an HTML document or is it part of the directory entry or is it in some system database?
@Robert: ok, it seems that you have noticed that the dialog still lacks a little from a usability point of view 🙂
So, please consider that the first aim during this GSoC has been make every piece and every list model work together both with Nepomuk backend and (now) with KFileDialog, I must admit that sometimes the UI has been put aside; but I promise it will receive some love in the next few days/weeks.
Metadata are all those informations which describe not only the document itself (and its content), but also its relationships with other documents, Web sites, world places, people (from your address book) and so on… and they are part of Nepomuk, so they are saved by the Nepomuk backend in a system db, like the informations that Strigi extracts from your files.
OK, I’ll leave it your capable hands.