You are viewing an old version of this page. View the current version.
Compare with Current
View Page History
« Previous
Version 3
Next »
- ITInfo
- XML Query
- Not able to query XML to generate index pages (no XQuery), either at author time or runtime
- Detail: We initially planned to have all of the current ITInfo content migrated to XML. We would then use something like Xquery to query the XML, gather the appropriate data, and display said content to a page. We have recently discovered that there is no way to effectively query XML content in this fashion, so we are investigating work-arounds.
- More AVM in XSL
- Almost all MIT author-time work-arounds require author time extensions via XSL which cannot trigger all of the events that we need; need more AVM functions in XSL
- Detail: "Our approach to adding functionality to the authoring process has been to use the XSL to make AVM function calls. However, we cannot get the XSL to execute in all cases where it is needed in order to have a consistent behavior. The repository has numerous events which can trigger some code to run (for example, to regenerate a new index page when a file in a folder is changed), but the WCM does not make them all available to us. Also, only a few AVM functions are ""built-in"" to the XSL and they are not enough for us; we have to create our own Java classes in order to access these functions. The whole AVM API should be available to XSL."
- Runtime DB cannot preview in Virtualization Server
- Any runtime db/app solution is un-previewable thru virtual server; might need different solutions at author time & runtime
- Detail: Solutions that use external tools cannot necessarily be run in Preview mode. If the data is stored differently at author time (in the Alfresco repository) or runtime (in file system, or XML database), then preview has to do something different in order to look the way it will look at runtime, and it may not be possible to get an exact correlation.
- Static Content Migration
- Embedding HTML / Unescaped Markup
- Difficulty with embedding HTML in the "generic" webform using TinyMCE
- Detail: We have a concern that our users are used to editing in Dreamweaver, and will not find the form-based GUI very easy to use. The usability (or lack thereof) of the editor should not drive users away from the application. Could be solved as of 2.0.1.
- Content Regeneration
- Regeneration of content when needed (see dependency mgmt)
- The import process involves a great deal of re-formatting the incoming content, going from HTML to templatized XML. Even with scripts and Java classes, we keep discovering many small issues around this.
- Workflow
- Workflow for non-Alfresco users
- Non-WCM users need to be included in workflow (email w/URL, click to approve/reject), see Authentication for access issues
- Detail: The workflow must account for non-Alfresco users. A non-Alfresco user must be able to view content in question and possibly comment on the content as well.
- Commenting
- Need commenting on tasks during workflow
- Detail: Commenting on content must be part of the workflow, rather than simply offering a reviewer the option of approving or rejecting.
- Task List
- Need list of files in a workflow task
- The same users will probably be part of many workflow processes going on simultaneously (there are only 2 final reviewers before content is published). These users need to know what files a particular task refers to
- Workflow Override
- Need originator/admin to be able to expire, cancel, and move workflow along without waiting for completion
- Detail: Sometimes a reviewer may not respond in time, or a piece of content has to be elevated in urgency. Being "in flight" means a content item cannot be edited until the workflow is complete, and it will be necessary to have the ability to override the process.
- Pooled Workflow
- Pooling workflow has especially poor UI
- Detail: We've tried workarounds to make parallel workflow act like pooled workflow, to no avail.
- Authentication (Authoring Environment)
- Authoring time authentication done via Acegi plugin; Alfresco uses a very old version of Acegi
- Detail: We are in the process of standardizing on Acegi within ISDA; we are developing our own authentication implementation of Acegi and custom decision mangers... can we use our Acegi implementations to provide the authentication and authorization for Alfresco.
- Deployment
- v2.0.1 only deploys to runtime Alfresco repository, we want flat files for web server when possible
- Detail: MIT requires deployment to a file server
- Link Management
- Basic Link Validation
- Links need to be validated
- Detail: MIT is aware that Alfresco has developers working on this piece, although there has been little communication from Alfresco about what exactly they are building.
- Automatic Link Maintenance
- Need links to automatically update if a file is moved to a different location in the web project. All links that were associated to the old location should regenerate and reassociate with the new content location.
- Detail:
- Search for and create links while editing content
- Need to easily search for existing content and create links while editing content pages
- Detail:
- Dependency Management
- Our solution requires our own build of the Alfresco WCM client, and there are some issues with respect to the association between source and generated content.
- Detail: Because there is no dependency mgmt, when an XSL for an Xform is changed, the files that were generated by that XSL are not updated. We have written our own Java class and added a button to the WCM client to regenerate all files so as to get up-to-date HTML from our XML. Because of this, we have to compile & build the WCM client ourselves in order to add our new widget. We cannot deploy an out-of-the-box Alfresco distribution. It does not always work perfectly, because the association between a generated HTML and the XML/XSL that it came from is not as tight as it should be – there are cases where the HTML has to be manually deleted.
- Application and Configuration Management
- Full server restarts for configurations
- Too many configurations are internal to the web application and require a full restart
- Detail: "Examples of internal configurations to the web application require a restart:
- Workflow changes
- TinyMCE reconfiguration"
- Virtualization server performance issues
- Virtualization server gets wedged for no clear reason & requires re-start; or it doesn't work, and then spontaneously starts working for no known reason.
- Detail:
- Instance restoration
- Cannot easily restore an Alf instance
- Detail: Exporting and importing a repository does not appear to be sufficient to get back to the original state.
- Clustering WCM servers
- Cannot cluster WCM - this is important for all of our projects that plan to work with the Alfresco repository
- Detail: "MIT needs a way to host their authoring environment in a clustered environment that provides failover protection. Nightly backups of content could be a temporary fix, but we would prefer to have a more sophisticated set up. It would be best to not put our users at risk of losing a days work if the server crashes. In addition, we cannot afford to lose information for clients that exist outside of IS&T (such as the MIT Alumni Association, among others)"
- Hosted service capabilities
- Does not handle a "hosted service (with multiple separate and distinct websites/web projects)" environment well. The "Company Space" is shared by all, along w/categories, dictionary, etc. We want to domain-ize different user groups w/different apps/projects, ideally each has a Company Space. - this is important for all of our projects that plan to work with the Alfresco repository
- Detail: "We are not going to move forward with a ""one Alfresco instance per website"" setup. We need a way to host separate and distinct web projects on one Alfresco instance, including separation between content, users, roles, etc. In addition, we would like more sophisticated user management within specific web projects. For instance, we would like certain users to hold the role of a ""reviewer"" for most of the site but also hold the role of ""content manager"" within a specific section of the website."