open questions for web api publishers

I think web style software is still a kind of a new thing. People are
starting to get it, but where are the tools to facilitate its development?
So, here is a list of questions open for any web api publisher to answer.
This includes publishers like amazon, evdb.com, delicious, google, upcoming,
technorati… the list goes on.

  1. How was the URI scheme decided upon?
  2. What are some aspects of the URI scheme that really worked, or you
    thought could be better?
  3. Which tools, if any, facilitated exposing your business and
    persistence logic through the URIs?
  4. Why and How did you decide on the format that represents your
    application’s information?
  5. What alternatives to the current XML format were considered?
  6. Are there any considerations to add a “help” and “test” or
    “advise” output formats? (a help format might facilitate automatic
    discovery or dynamic help in a client application). (test or advise
    might also facilitate automated discovery and consumption of services
    as well as allow practice getting the api’s consumption correct by
    explicit sanity checking, no-ops, and random data).
  7. How does your authentication model work? What decisions most affected its
    design?
  8. If there were money attached to usage of the API, how might that
    affect your authentication scheme?
  9. If the API allows developers to set up their applications to act as
    proxies for end users (who have money attached to their accounts, how
    is authentication affected by that requirement?
  10. Do you support OpenId? Why not?

If you are writing software that works with the internet, this knowledge
should be a part of your toolbox. It should be as common as “how arrays work”
or “how pointers work.”

Advertisements

Post a Comment

Required fields are marked *
*
*

%d bloggers like this: