Kendra Base Wish List
- When a user is not logged in what are their default relationship and search servers.
- Multiple search instances on make query pages.
- Remove "User:" from all URL references pointing to users in:
- Error Checking:
Need to check for slashes at the end of the Relationship Servers and Search Servers URLs. If not then add 'em. Done.
Need to trap when no file selected for upload csv. Done.
- Menu Sidebar:
I've got a question mark over the userid being given to the user. I don't think it's information they need to know, no? Done.
The user page and talk page - I don't see how they fit into the Kendra Base model. We've got it down for future development that any page can be owned by an author and the author can then admin access rights. So, having these pages kind of fudges that idea, no? Can we discuss/hide these menu item pages for the moment? Done.
Make global recent changes work Done.
- Make "what links here" do subj + pred + obj all at the same time
Change menu items to read: Relationship Servers (not trusted) and Search Servers Done.
- Consistent presentation of text in urls and pages URLs:
- Spaces (and other characters) in URLs and property drop down menus:
In 'Make query' mode the property pull down menu has underscores where, I think, there should be blank spaces - just for the display though. Done.
- In the results page can all "_" in the text (not the URLs) be converted to spaces?
In left menu column 'Main page' link is: http://base2.kendra.org.uk/kendrabase/main%20page - as do most other menu links actually. Could it just be an underscore rather than HTTP character for space? This has been done but I'm leaving this item in for reference! ;-)
- The colon in the user pages needs to be displayed as a colon in the URL.
- Just for consistency... can we have all system generated URLs be in lowercase?... So...
- User -> user
- Upload_CSV -> upload_csv
- Upload CSV:
- when uploading user needs to be asked what class is this linked to and what in the prefix for the page names that are created. Then the uploader automatically generates the full page names by taking the prefix and adding a id number to make them unique. Also, the user is presented with a list of class by doing a search for all the defined classes in the system. So that means there will be 3 input fields: file, prefix and class.
- Semi-colon ";" (or user defined de-limiter) separated file input form for getting data into KendraBase.
- Upload file form. Headings in first line.
- These are the headings you have chosen (with sample preview?). OK?
- What class are you inputting instances of?
- User decides: What field acts as suggested page-name? Unique?
- File to input?
- Separator to delimit fields?
- Test file?
- Output as CSV so we can 'complete the loop'.
Use POST instead of GET? "Save this search"? Or a compressed URL?
- Savable Queries:
- Give option for 'self' that would allow for form (with all values intact) to be replicated from URL.
- Also, allow query results page to be replicated from a URL.
- New Server Names:
- Need to depreciate all the dev#.kendra... names soon.
- Edit As Form:
- Where we can modify an instance/page using a form as we input an instance.
- Name/value no class stuff. So we just take the page and turn it into a form where the properties are hardwired (not drop downs) and the values are presentent in text input boxes.
- Seems to be lost in http://base4.kendra.org.uk but is still in http://base3.kendra.org.uk.
- Can we get it back into the main code?
- NH comment: what's going on here?
- rm "string,lower" typo - not sure what this does, Neil?
-- FIXED: simply a typo, wasn't exercised by most code paths, now fixed
- Make Query:
- Add blank option to auto-generated pull-down lists and make them work again
--- DONE: but need to make back-end code work again...
- Add special-purpose numeric matchers to list of operators - so full list should be literally:
- number equal to
- number greater than
- number less than
- number not equal to
- text equal to
- text contains
- text does not contain
- text starts with
- text ends with
- text greater than
- text less than
- text not equal to
-- DONE: but need to make back-end code work again...
- User Server Pages: Trusted and Server:
- Need to review the situation -- and verify that it still works after changes.
- Every user has their own:
- Trusted server list page
- List of page URLs that contain dictionaries A is_the_same_as B that will be included in a user's search.
- Search server list page
- List of server URLs that will be included in a user's search.
- Needs to have Real Name of participant. See Account Management above.
Thinned from First Stage:
- Table (columns and rows) display of results for query. Moved the following:
- Check boxes: SHOW: include in output
- Radio button: Primary sort field
- Limit to certain number of results
- Optional paging given enough time.
- Lock editing of pages.
- Site owner edit anything
- Pages locked by users with "locked by: person" property - in band. Read only.
- Invisible pages
- Pages made invisible by users with "invisible by: person" property - in band. Private.
- When editing a specific object's properties, if the object belongs to a specific object type (like song performance), then every object property should be in buttons so users can easily know and quickly enter any qualities suitable for this object type
- Enable integration with other user systems:
- Import: blogs, address book, calander, messaging...
- Fine grained permissions: Read only and invisible fields
- Class inheritence of permissions
- Forms (Input and Query) more customisable.
- Storing BLOBS - images and sound files.
- Directory tree mapping
- Rewrite underlying database to address objects by ID and not by name. So that we can differentiate between ID and display name. And it'll probably be faster to operate. And save space.
- Templating and transclusion.
- film-catalogue with searchable meta-data which connects log-files with the original material (to the original timecode)
Really Future Ideas: