Introduction

This is a bag of issues around Longwell, see also Banach, Fresnel Todo, Priority Codes.

Issues

• P3 Fresnel stack traces (see below)

• P4 Processing of file loading is incomplete (see below)
• P4 longwell-csi configurations are a fork from the original, better, retire original?
• P4 The drop down view arrow is displayed even if no other views are provided; adapt the patch that adds it to hide it and only show it if other views are available.
• P4 The grid view has certain minor issues to resolve: the grid is really expecting thumbnails smaller than a certain size (via CSS), the configuration might be nice to do via a dialog in addition to the configuration.vt method. The code may need some vetting to get rid of unnecessary code revolving around the 'column properties' that came in from imitating the tabular view.
• P4 the various longwell-csi views should be configurable, meaning that longwell-csi should not load them by default (along with all the required javascript files) if they are not used
• P4 it is not currently possible to pre-configure the timeline and map views to avoid having the user to select which property to add.
• P4 The facet boxes should have an enlarging handle like we do for Exhibit; this feature in Exhibit comes out of the Ajax toolkit - should we incorporate the entire toolkit for this feature (and others where it makes sense), try to extract just this resizing feature, or leave this on the wishlist?
• P4 It would be very cool if one could load data to longwell-csi into a 'configure' mode and then use its own UI to configure it, for example to select which facets to hide, which columns to show, what order to default for the items, what properties to use for timelines and maps, and so on. As things are right now, the configuration is done via javascript in the embedding web page, but it would be nice if such javascript text could be given to you by longwell itself when you are done with your configuration, so that you can cut paste it into your own page and be done with it.
• P4 Piggy Bank should be able to get data out of longwell-csi pages
• P4 The Dwell<->server protocol needs a more formal/elegant/extensible design
• P4 The template language used in exhibit is accumulating an installed base, so supporting it might be nice, doc would be good at that point.
• P4 badly formed URI raises Javascript alert
• P4 refine algorithm that determines when identifiers are links: now many non-resolvable (spurious) links created in lenses or list views
• P4 devise a way of handling facet name collisions - i.e. distinct facets both labeled (e.g.) 'thumbnail'
• P4 when one executes a search and receives differently typed results, choosing the 'missing' value in a facet that one finds only on one type will eliminate that facet since it's no longer a common property of the newly restricted results - we don't quite like this

• P8 popup view on facet values?
• P8 long strings negatively effect longwell UI, find way to mitigate
• P8 add a new parcel for displaying current restrictions which doubles as a place to remove restrictions
• P8 language bucketing
• P8 Migration path for Timeline/Exhibit users into Longwell

Design Notes

• We like configurations, they reduce forking
• Sesame2 is good.
• Sesame2's memory story is nice, compact and fast
• Stacking sails is nice.
• Why is the disk store 4-5 times slower than the memory story?

Current Bugs

Bug 2

sometimes a stacktrace like this is found. the weird thing is that it's always 6144!!

java.lang.ArrayIndexOutOfBoundsException: 6144
        at org.mortbay.jetty.AbstractGenerator$OutputWriter.write(AbstractGenerator.java:766)
        at java.io.PrintWriter.write(PrintWriter.java:354)
        at java.io.PrintWriter.write(PrintWriter.java:371)
        at org.apache.velocity.runtime.parser.node.ASTText.render(ASTText.java:61)
        at org.apache.velocity.runtime.parser.node.ASTBlock.render(ASTBlock.java:55)
        at org.apache.velocity.runtime.directive.Foreach.render(Foreach.java:166)
        at org.apache.velocity.runtime.parser.node.ASTDirective.render(ASTDirective.java:114)
        at org.apache.velocity.runtime.parser.node.ASTBlock.render(ASTBlock.java:55)
        at org.apache.velocity.runtime.parser.node.ASTIfStatement.render(ASTIfStatement.java:70)
        at org.apache.velocity.runtime.parser.node.ASTBlock.render(ASTBlock.java:55)
        at org.apache.velocity.runtime.directive.Foreach.render(Foreach.java:166)
        at org.apache.velocity.runtime.parser.node.ASTDirective.render(ASTDirective.java:114)
        at org.apache.velocity.runtime.parser.node.SimpleNode.render(SimpleNode.java:230)
        at org.apache.velocity.Template.merge(Template.java:256)
        at org.apache.velocity.app.VelocityEngine.mergeTemplate(VelocityEngine.java:450)
        at org.apache.velocity.app.VelocityEngine.mergeTemplate(VelocityEngine.java:419)
        at edu.mit.simile.longwell.command.APICommand.execute(APICommand.java:44)
        at edu.mit.simile.longwell.LongwellServlet.doPost(LongwellServlet.java:516)

This error is caused by a jetty bug, that is fixed in 6.1.2rc1. See http://jira.codehaus.org/browse/JETTY-244 for more information.

Bug3

This produces a wealth of uninformative errors dealing with Fresnel:

ERROR:  ''
ERROR:  'java.lang.NullPointerException'

one per result item.