On 28/06/12 13:06, Dave Abrahams wrote:
This ship has undoubtedly sailed, but I'd like to register my displeasure with names like "uniqued," "taken," and "strided" et. al. The effect (subjectively of course) is to take a beautiful, minimal abstract notation that reflects a common practice (seriously, it's "grep," which is a verb, not "grepped") and mess it up with unconventional and awkward linguistic insurance against misunderstanding.
this-has-been-a-test-of-the-emergency-nattering-system-ly y'rs,
Dave, This came up for discussion during the review of Boost.RangeEx. Ultimately the majority preferred the naming and syntax we currently have. While acknowledging the importance of finding clear names that identify our abstractions, the choices here are both equal in this respect. To my estimation there is no impact on external quality factors and therefore while there will be strong opinions the outcome of this debate does not appreciably affect the quality of the library. Both naming conventions are unlikely to create confusion. I don't recall your input during the review on this matter. There are clearly some adjustments that can be made: 1. I could add the non-'ed' suffixed versions in a manner that minimizes backward compatibility issues; 2. We could revisit the idea of having both the function and pipe forms; 3. We could provide a new namespace with new names. I am very open to changing these especially since I believe we can do this without negatively impacting my treasured early adopters. Your displeasure troubles me. I am sure you have some ideas of actions you would like taken to alleviate the displeasure. Would you mind being a little more explicit about what those might be. Please remember that I'm of limited intellect, please use small words! Thanks, Neil Groves