In the context of this article, a Triaster Server's identity could be its computer name or an alias. Perhaps the most obvious manifestation is in the addresses of the process library websites, e.g.
http://processlibraries.triaster.co.uk/processlibraries2011/triaster sample library
Some explicit references to a host are necessary. For example, an alert e-mail that contains a link to library content would need that link to be fully-qualified. Any item that is outside of a Triaster Server but needs to link in would require it; items that reside within a Triaster Server shouldn't.
but would be better expressed as:
Note the preceding forward-slash, which denotes the path is relative to the root of the website. Not only is this URL independent of the host identity, it would remain valid if the protocol were to change from HTTP to HTTPS. A slight disadvantage is that such a link is unlikely to work from a Visio file, but that's only relevant when developing maps, not when they are published.
When changing a server's identity, these are the configuration changes that would need to be made.
This is the main configuration file, and any changes should be preceded by taking a copy that could be restored.
This is the server's identity as a Publication Server as seen by Triaster Server's Token Server service, and should correspond to an entry in the Token Server's Server List.
It doesn't have to correspond to a server's computer name, but for consistency and clarity, it should. This needn't change if just adopting the use of an alias for website URLs.
This should correspond to the 'ServerName' above.
The fully-qualified URLs are pertinent to the construction of links from e-mails to a process library. These should change whether changing the computer name or using an alias.
URLs need to change.
There are unlikely to be any host-specific references, particularly in later implementations, but that should be checked.
These may contain explicit references, but needn't. So in the simplest case of adopting the use of an alias, all that should need changing are the site-scoped URL settings in 'Settings.xml' and host references in 'Libraries.txt'.
By restricting the use of a Triaster Server's identity to things that work from outside a Triaster Server, content within is made more portable, and the identity itself is easier to manage.