Re: blocking communication?
"Henrik I. Christensen" <[email protected]>
| Newsgroups | gmane.science.robotics.orocos.user |
|---|---|
| Message-ID | <[email protected]> |
>>>>> "WP" == Wolfgang Ponweiser <[email protected]> writes: Hi! WP> Why do you use blocking communication patterns? My WP> answer: 1) I know that design and implementation of WP> non-blocking communication produces a bit more effort. 2) But WP> permitting only non-blocking communication can avoid the WP> dangerous situation of reciprocative blocking and hence WP> deadlocks. There is nothing in the current communication patterns that prevents non-blocking communication. You send of a query and subsequently you are notified of the result! This is exactly way non blocking communication is carried out. The blocking patterns a bit more developed and can be used for synchronisation (that is often a side-effect). So there is NO conflict here! WP> Isn't it better to use only non-blocking communication, WP> because it enables a seperation of communication and WP> configuration? (Seperation of communication and configuration WP> = allows using differnet tools, seperated development, ....) The is NO need for general configuration to be mixed with communication this is a mixup. There might be a need to configure *communication* / data analysis but for general "module"/component configuration there is no need to tie this to communication (that would actually be a disadvantage), as we might want to use other mechanisms such as configuration files ... for this! Henrik --------------------------------------------------------- Henrik Iskov Christensen Phone: +46 8790 6792 Professor, Director Fax: +46 8723 0302 Centre for Autonomous Systems E-mail: [email protected] CVAP/NADA GSM: +46 70 670 6240 Kungliga Tekniska Hoegskolan URL: www.nada.kth.se/~hic SE-100 44 Stockholm, Sweden --------------------------------------------------------- Delivery address. Fiskartorpsvägen 15A, 10044 Stockholm ---------------------------------------------------------