| 2011-03-03 09:21:38 utc | jmettraux | tosch_le: hello thanks again for yesterday |
| 2011-03-03 09:22:26 utc | tosch_le | never mind, i have to thank you. |
| 2011-03-03 09:22:38 utc | tosch_le | thanks for releasing the fixed gem :-) |
| 2011-03-03 09:28:39 utc | jmettraux | it was still broken : I had forgotten public/ and views/ |
| 2011-03-03 09:30:31 utc | tosch_le | should have noticed that, too. the old gem versions in the README and the rails template were embarrassing, too. |
| 2011-03-03 09:30:56 utc | tosch_le | i'm glad we have such a nice community :-) |
| 2011-03-03 09:32:03 utc | jmettraux | :) |
| 2011-03-03 10:16:14 utc | jmettraux | why are [ruby] devs always equating workflow to state machines ? Though they say my "photo workflow" or "my git workflow" but never think of it as a state-machine-esque thing |
| 2011-03-03 10:21:19 utc | tosch_le | you are asking deep question, you are… |
| 2011-03-03 10:23:42 utc | tosch_le | could it be that state machines are so prominent in computer sience lectures? |
| 2011-03-03 10:25:43 utc | jmettraux | maybe |
| 2011-03-03 10:26:51 utc | tosch_le | another point is that many workflows may be implemented by using a state machine. |
| 2011-03-03 10:27:08 utc | jmettraux | yes |
| 2011-03-03 10:27:59 utc | jmettraux | ACTION digesting Torsten's lines |
| 2011-03-03 10:27:59 utc | tosch_le | and in the end, human language is ambigous: there is no clear definition for the term "workflow" |
| 2011-03-03 10:28:11 utc | jmettraux | whereas state machine is well defined |
| 2011-03-03 10:29:04 utc | tosch_le | ACTION has the impression that devs like well defined things and dislike ambigousness. |
| 2011-03-03 10:31:48 utc | jmettraux | for the record http://ruote.s3.amazonaws.com/mindmap.pdf |
| 2011-03-03 10:36:01 utc | jmettraux | tosch_le: so forcing workflow into state machines, could it be that devs try to limit the workflows when modeling them with non devs ? |
| 2011-03-03 10:36:33 utc | tosch_le | i suppose that could be right, yes. |
| 2011-03-03 10:37:28 utc | jmettraux | for you, what's the difference between a state machine lib and a workflow engine lib ? |
| 2011-03-03 10:37:34 utc | tosch_le | devs know state machines and their logic, they can handle them easily. most of them don't really know about workflows or business logic, they try to simplify them in order to understand |
| 2011-03-03 10:37:43 utc | tosch_le | or better: handle them |
| 2011-03-03 10:38:17 utc | jmettraux | ok |
| 2011-03-03 10:39:13 utc | tosch_le | short answer: the difference lies in the names: a state machine lib gives you tools to implement a state machine. a workflow engine [lib] gives you tools to drive workflows. |
| 2011-03-03 10:39:35 utc | tosch_le | so for me the difference is between state machine and workflow |
| 2011-03-03 10:40:05 utc | jmettraux | there is a overlap for, let's say, simple workflows |
| 2011-03-03 10:40:23 utc | jmettraux | when does the overlap stop ? |
| 2011-03-03 10:40:43 utc | tosch_le | let's not call it overlap. there are different ways to implement a workflow, a state machine may be one of them. |
| 2011-03-03 10:41:03 utc | tosch_le | both are different levels of abstraction imho |
| 2011-03-03 10:41:34 utc | tosch_le | a workflow is about business logic, a state machine about program/software logic |
| 2011-03-03 10:43:15 utc | tosch_le | some workflows may be implemented by using a state machine, some not. |
| 2011-03-03 10:43:29 utc | jmettraux | when in a computer, business logic is symbolized by software logic |
| 2011-03-03 10:43:47 utc | jmettraux | an invoice is symbolized/represented by a data structure |
| 2011-03-03 10:43:49 utc | tosch_le | yes, that's what i mean: different levels of abstraction |
| 2011-03-03 10:44:20 utc | tosch_le | the workflow itself has to be abstracted into a program logic |
| 2011-03-03 10:44:30 utc | jmettraux | ok |
| 2011-03-03 10:44:59 utc | tosch_le | that can be done by using a state machine, but also using an "operating system for business processes" ;-) |
| 2011-03-03 10:45:05 utc | tosch_le | i like that claim for ruote |
| 2011-03-03 10:45:24 utc | jmettraux | :-) |
| 2011-03-03 10:46:07 utc | jmettraux | speaking of operating systems and the programs it hosts |
| 2011-03-03 10:46:20 utc | jmettraux | this is the most brilliant workflow I've seen these days : http://teddziuba.com/2011/02/stupid-unix-tricks-workflow-control-with-gnu-make.html |
| 2011-03-03 10:47:26 utc | jmettraux | workflow : operating system for business processes / state machine : attaching callbacks to files |
| 2011-03-03 10:49:44 utc | jmettraux | running shoes on, I'll be back, many thanks for the conversation ! |
| 2011-03-03 12:28:50 utc | jmettraux | back |
| 2011-03-03 12:37:24 utc | tosch_le | welcome back |