[chef] Re: Cookbook statistics?

[Anton Baranov < >]

Hello, Brad.

Not sure, that this is possible. There are some reasons for that.

1) Process of rewriting bash-scripts into set of chef-receipts/cookbooks more depends on the logic of bash-scripts, than of the count of lines.

2) Estimated time for writing all needed cookbooks should be calculated depending on the exact number of changes cookbooks must produce on the server.

-- 

With best regards, Anton Baranov.

вторник, 10 сентября 2013 г. в 0:23, Brad Knowles написал:

Folks,

I'm curious to know what kinds of statistics you track regarding your cookbooks and your cookbook development. Do you track how many lines, words, and/or characters you have on a per cookbook or recipe basis?

Do you do anything else to try to compare and contrast your cookbook development versus more traditional development languages and methods?

Do you do any kind of estimation of effort before starting actual development of your cookbooks?

I've been involved in converting a large installation shell script into a collection of recipes inside of a cookbook, and even if you cargo-cult the bulk of the script as-is, this process has still taken a lot longer than the customer expected, and certainly longer than I would have liked.

There was clearly a disconnect and miscommunications there, but I'm now trying to better understand underlying metrics and methodologies that I should be using to try to prevent these kinds of happening again in the future.

Should I be counting the number of lines of code, or words, or characters in the shell script and using that as a basis to try and give some sort of indication how long it might take to convert?

Is there a typical ratio of lines of code versus words versus characters that I should expect to see?

Or am I looking at the wrong end of the telescope?

Thanks!

--

Brad Knowles < "> >

LinkedIn Profile: <http://tinyurl.com/y8kpxu>