aboutsummaryrefslogtreecommitdiffstats
path: root/doc/state-of-code.md
diff options
context:
space:
mode:
authorMufaddal Makati <mmakati@mmakatimbpr15.cablelabs.com>2015-08-07 13:09:49 -0700
committerMufaddal Makati <mmakati@mmakatimbpr15.cablelabs.com>2015-08-07 13:09:49 -0700
commited73dbf357aff41edcbab401a94e5fbc266d9391 (patch)
treee3561a9290e4bd0de19704228867ec91910e68ed /doc/state-of-code.md
parentf44f4aa8b5eb244fc7603c17e875eebf0f3b0b67 (diff)
First Commit
Diffstat (limited to 'doc/state-of-code.md')
-rw-r--r--doc/state-of-code.md22
1 files changed, 22 insertions, 0 deletions
diff --git a/doc/state-of-code.md b/doc/state-of-code.md
new file mode 100644
index 0000000..0a9eff3
--- /dev/null
+++ b/doc/state-of-code.md
@@ -0,0 +1,22 @@
+This is a proof of concept prototype implementation enabling the creation and management of MEF EPL Services.
+
+Being a proof of concept prototype, there are some areas which have received only superficial coverage or no coverage at all. Following are items that should be addressed when evolving this to a production oriented implementation
+
+1. The current test suite is reasonable, but exhaustive. Production targeted version should have more thourough test coverage
+2. Exception handling is only superficially implemented. A production system should include a more robust exception handling and error reporting architecture
+3. Only a subset of the MEF Attributes have been implemented. Long term the goal is to evolve this code base to fully support all MEF paramaters
+4. We are only implementing a portion of the data model for the sake of expediency.
+ - Only EPL service is supported at the moment
+ - Currently, the service repo is hard coded to hold EPL (we have not implemented the more general MEFSvc and Eline/ELine/Tree classes that in the future will also reside in the repo).
+ - The MEF service hierarchy (MEFSvc->Eline->EPL) need to be architected and implemented
+ - The SvcMgr needs to be architected in order to handle multiple service types from a REST API perspective and from a persitance perspective
+ - the REST URL/path architecture is currently hard coded, should be configurable at some point (config file, etc).
+5. There has not been a definitive analysis of which specific service attributes are recieved, generated, or queried at each service layer.
+ - It would be very useful to do this analysis, and then reflect in the code through structure and documentation, which attributes are set/queried at each layer
+
+Specific To Do:
+1. Who supplies/generates evcMaxSvcFrameSize? Currently hard coded to 1600 in EplService.create()
+2. Make sure for all XXXClient.create(xxx), that we capture the returned version, in case the service layer modified it
+ - for example epl = eplClient.create(epl)
+3. Check for nulls when prior to printing out lists in dump() fxns
+4. Unique EVC IDs are being generated via an internal counter. This of course does not scale nor survive restart. A proper unique ID generation approach will be needed for production. \ No newline at end of file