Generic functions in TurboGears
Download as PPT, PDF1 like389 views
This document discusses generic functions, which allow functions to be selected based on the context or type of arguments passed. It provides examples of using generic functions in TurboGears to make frameworks more extensible and data-driven. Finally, it discusses design considerations for generic functions like limiting introspection and avoiding cryptic errors.
![Generic f unction s in Simon Belak [email_address] EuroPython 2006](https://image.slidesharecdn.com/2605902/85/Generic-functions-in-TurboGears-1-320.jpg)
Generic functions in TurboGears
- 1. Generic f unction s in Simon Belak [email_address] EuroPython 2006
- 2. What those not attending will miss Overview of generic functions Practical examples in TurboGears Generic functions from a framework designer’s point of view
- 3. A strange beast by the name of generic function
- 4. Generalisation of message passing Problem: a set of abstractly the same, practically different functions Solution: choose based on context (e.g. arguments)
- 5. All for one, not one for all Single dispatch Python, Java Multiple dispatch Type based CLOS, PyPy, Python 3000? Identity based CLOS Predicate based Cecil, Mathematica, RuleDispatch
- 6. Pick me, pick me! Relation more specific Arbitrary dispatch strategies
- 7. Combine and Conquer Where is my super? method chaining (dispatch strategy applies) How is this better than sliced bread? before (all applicable) after (dispatch strategy applies) around (all applicable)
- 8. Something is rotten in the package of RuleDispatch Limited introspection Magic Scarce documentation Cryptic errors Once a generic function, always a generic function
- 9. Call now and get all these wonderful buzzwords free of charge DRY Ad-hoc event/plug-in system End-user framework extensions Data driven
- 11. Use-case #1: JSON for everyone Data driven End-user framework extension
- 12. Use-case #2: errors bring joy to my heart Data driven Ad-hoc event system DRY (user and framework)
- 13. Use-case #3: not all see me equally exposed Data driven DRY
- 14. A word to the wise
- 15. RuleDispatch pitfalls Compile on definition (lexical scoping) Constant folding Solution: Limited inference
- 16. Hide but do not lock away Rule generation Hides underlying complexity Users need not know: solve the most common case implicitly Registrator pattern Adjust the scope Always leave an out
- 17. The user takes precedence Custom dispatch strategies
- 18. Many words, few lines
- 19. The future is bright, the future is generic PEAK-Rules Python 3000?
- 20. Questions?
- 21. References peak.telecommunity.com/DevCenter/CombiningResults www.gigamonkeys.com/book/object-reorientation-generic-functions.html www.python.org/pycon/2005/papers/53/img0.html www.artima.com/weblogs/viewpost.jsp?thread=155123 codespeak.net/pypy/dist/pypy/doc/theory.html#multimethods http://citeseer.ist.psu.edu/ernst98predicate.html http://citeseer.ist.psu.edu/chambers99efficient.html