View of /PEAK/TODO.txt
Parent Directory
| Revision Log
Revision:
1191 -
(
download)
Wed Jun 18 18:27:39 2003 UTC (20 years, 10 months ago) by
pje
File size: 4998 byte(s)
All 'peak.binding' APIs now only accept positional parameters for items
unique to that API. Items common to multiple APIs (such as 'offerAs',
'doc', 'attrName', etc.) should now be supplied as keyword arguments.
Bindings also now automatically "suggest" the containing object as a parent
component for the contained object, whenever a value is assigned to them or
computed. If a non-None 'adaptTo' is set on the binding, the value assigned
or computed will be adapted to the specified protocol before the parent
component is suggested. 'binding.New()' no longer relies on the
'IComponentFactory' interface, but instead uses the new adapt/suggest
mechanisms.
Previously, parent components were only "suggested" when a binding was set
via component constructor keyword arguments. Now, this is done at any time
bindings are set, but *not* for non-binding keyword arguments. In other
words, ordinary attributes of a component do not receive "suggested parent"
notices, even when set via constructor keyword arguments. If you want an
attribute to do this, you must define the attribute with the binding API;
e.g. via 'requireBinding()' or 'binding.Constant()'.
Open Issues/To-Do Items
Targeted for 0.5 Alpha 3
* Refactor binding and naming frameworks to
- get rid of cruft (such as JNDI 'attrs' hangover) and duplication
- streamline descriptor mechanisms used by binding and model to make them
uniform and cleanly functional (in both senses of the word functional)
- clean up attribute metadata framework and get rid of all the special
cases (e.g. isinstance() checks, special attribute names, etc.)
- make indirection convenient (e.g. bind to looking up something from
another binding)
* Implement "contextual protocols" (c.f. "object teams") and "parameterized
protocols" (E.g. 'ListOf(IFoo)', 'MappingOf(keys=IBar,values=IBaz)')
* Rough-out web publishing framework
- Publication interfaces
- Viewer/Controller adaptation mechanisms
Targeted for 0.5 Beta 1
* Up-to-date reference docstrings for all packages
* Finish tutorial chapter 2 (?)
* Web publishing framework sufficient to deploy page-based or object-
published apps
Targeted for 0.5 Final Release (or sooner)
peak.storage
- unit tests for more complex object scenarios: references, thunks..?
peak.naming
- useful example "flat" naming context (e.g. like AppUtils.URLkeys)
- useful example hierarchical naming context (e.g. like JNDI's LDAP
context or a filesystem context)
- rework smtp: to return a factory object that supports open().
Also think about whether smtp should move elsewhere. Maybe
there should be peak.network or peak.internet for things like
smtp, ftp, etc contexts?
- add in Ty's cool "n2" (Namespace Navigator) shell for working with
naming providers.
Targeted for version 0.6
* Get XMI writing in place, w/transaction support
* Generate UML 1.5 and CWM 1.0 and 1.1, and add them to the
'setup.py' package lists.
* On-the-fly class combination (think "runtime module inheritance",
but possibly without the modules) for DMs.
* Lock management interfaces/API for peak.storage
* Make 'cluster' parser complain about things that would cause
the clusterit tools to choke or barf on the file, or which would
produce ambiguous or unintended results. (Or replace with ZConfig
schema that can generate clusterit files. And/or replace clusterit
tools with PEAK ones.)
Future Releases
(Note: some of the below is held-over from TransWarp and may no longer be
relevant as written, they are being kept on this list as placeholders for
ideas or problem areas that may need to be re-considered in future.)
Miscellaneous
* Functional tests that access "real" databases, LDAP, etc.
Simulator/Module Inheritance
* Allow 'declareModule()' to bootstrap non-existent modules; this might
let us create "virtual packages" made by assembling other packages and
modules.
* Need a strategy for handling "del" operations; they are currently
untrapped. This might be okay under most circumstances, but need to
consider edge cases.
* 'makeClass()' should probably become part of the core API, where
it can be used to resolve __metaclass__ conflicts during the first
pass of importing a module (prior to running 'setupModule()')
Messaging/ObjectSpaces
* Support for sending and receiving remote cache invalidation
messages between DataManagers.
peak.model
* Implement WarpCORE-oriented structural model, w/Querying support
* "Indexed" version of in-memory model?
Queries
* Refactor to use interfaces, if appropriate
* Incorporate into AbstractModel?
- Pros:
* Queries always available
* Each StructuralModel implementation can easily include its own
performance-tuned version of the basic items.
- Con: default implementation doesn't perform well on large datasets
* How much of framework needs extensibility? Should the predicate
classes be placed in the StructuralModel's namespace so that predicates
have their meaning assigned by the StructuralModel implementation?
peak.metamodels.uml
* Helper methods in Elements & Services for marshalling, common queries, etc.
* Generator framework
- Tagged values in stereotypes vs. main values?
- Should tagged values be copied directly into templates? Treated as
Python expressions?
- Should Services be generated using an Element class' "static"
(class-scope) methods/attributes?
- Are association-ends scoped?
- Would it be better to seperate them?
- What determines whether an implemented Service actually stores objects
or delegates this to its subclass services?
* Simple Zope product demo (upload XMI, then browse the model via the web)