I'll clean this one up, too.
My objective will be to
- inform users unambiguously about mistakes and how to correct them,
- recover useful features,
- be reticent w.r.t. deprecated features (duration!),
- document the result
-W
PS.: The December 6 issue of Time magazine contains the remarkable sentence: "...an invention [...] intended to simplify the [...] process by making it more complicated". Well, well, well,...
On 29/11/2010 22:31, Michael Neale wrote:One thing to remember is the Drools codebase is mature, like 8 years or something now. There are lots of half implemented things or semi implemented things that are not documented or intended for end user use :)I think that goes back to an attempt to let people use different words/languages without i18n - so probably a bad idea.
Unless people object, I propose getting rid of that behaviour and cleaning it up to be keywords - AS LONG AS if they put in a non valid keyword, the error shows a list of what *is* valid so they can then correct it.
Partly that's beacuse you start working on an idea and never get it finished, but the code still lives on as you always intend to come back to it, but you never do.
Mark
_______________________________________________ rules-dev mailing list rules-dev@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-dev
On Mon, Nov 29, 2010 at 11:53 PM, Wolfgang Laun <wolfgang.laun@gmail.com> wrote:
Hi Michael,
I just discovered ActionType.addNewActionType, where the code tries to follow some
(now) completely undocumented principle, where columns can be identified by single
letters, the first one of the action type. This is in conflict with the documentation,
where only full-fledged keywords are permitted.
There are some undocumented keywords, e.g., DESCRIPTION. Using this results
in a duration attribute (which is deprecated anyway). This can be fixed, but then
the description is entered "as is"; it should have a leading '#'.
Clear out all undocumented things? Or fix them properly and document them?
What is it to be?
-W
On 28 November 2010 23:33, Michael Neale <michael.neale@gmail.com> wrote:
nice work..yes "syntax cushioning" is the best term I have heard for this.
I am sure your enhancements would be welcome.
On Mon, Nov 29, 2010 at 5:14 AM, Wolfgang Laun <wolfgang.laun@gmail.com> wrote:
I have, at long last, overcome my disinclination against spreadsheets and played around a bit.
As one of the incentives (perhaps the main one) for this kind of rule authoring appears to be a "syntax" cushioning by spreadsheet entries, I feel that additional simplifications might be appreciated. Therefore, I have modified some classes in org.drools.decisiontable.parser, to achieve the following, in the area of RuleSet entries:
Opinions, please, and should I just release this, or would someone care to have a look and test it?
- All entries are now repeatable, either by adding more cells to the right of "import" than just one (with a comma-separated list) or by writing more that one "Import" row.
- Same for "Variables", "Functions" and "Queries".
- All tags ("Import",...) are case insensitive and immune against leading and trailing spaces.
- Some user errors don't cause NPE; they throw an exception with an explanatory message
-W
_______________________________________________
rules-dev mailing list
rules-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
--
Michael D Neale
home: www.michaelneale.net
blog: michaelneale.blogspot.com
_______________________________________________
rules-dev mailing list
rules-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
_______________________________________________
rules-dev mailing list
rules-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
--
Michael D Neale
home: www.michaelneale.net
blog: michaelneale.blogspot.com
_______________________________________________
rules-dev mailing list
rules-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev