Capturing Pancho's thoughts about a 7.x-2.x roadmap, which were sent via email:

You have seen my task to convert amazon to the Entity API. I'd better get this
done by my presentation in 10 days, because the flexibility and compatibility
this would offer is exactly what I need to connect Amazon data with the data
from the local bookstores. However I see that this is my personal interest and
not necessarily the most pressuring issues for the forthcoming release.
So I'm really okay if you say: "Go play with it in your sandbox (as I
currently do) and then file it against v7.x-2".

On the other hand, I'd call it a smooth conversion to change the internal
methods now, and then focus on a larger overhaul for v2.
Cause actually I'm not so sure about the feture sets of the different modules.
The table-based datastructure is IMHO certainly more than a pure API. A pure
API should retrieve the XML data from the net and interpret it.
The demarcation between amazon API and amazon_media also remained unclear to
me - has it just been an arbitrary decision to make the API save images and to
leave some type-specific fields (for books, music etc.) to amazon_media?

So IMHO we could actually split (step 1) what is today the main amazon module
into a real basic amazon API and an advanced item management module named e.g.
amazon_item.
The former, amazon API, would get merged (step 2) with amazon_search, which is
just a tiny support function for a core module. The formal dependency on
search would be waived - if search module is active, it get's supported.
The latter, amazon_item, would also incorporate (step 3) what is today
amazon_media and would become more flexible to cater all kinds of items and all
needs.
amazon_filter would still work with or w/o the amazon_item component, just as
amazon_field (aka asin) would. However I can see at least the latter being
integrated into the basic Amazon API.

Now let's take a look (step 4) at amazon_store. Sure, we're talking about a
big Internet store, so people just want to sell their stuff.
But wait: I can see plenty of use cases for searching and browsing the
categories and using browsenodes and stuff for other purposes. And I see cases
where the included cart is more a burden than anything else, because they need
something way more sophisticated.
So I can see some use in splitting amazon_store into something like
amazon_browse and a relatively shrinked module called amazon_cart, amazon_shop
or whatever.

Leaves it to 4 sub-modules: amazon API, amazon_item, amazon_browse and
amazon_shop, each catering a specific set of use cases. This would also mean
that the two projects had to become one, but I guess the separation has
historical reasons anyway (you wrote amazon_store and then took over the main
project).

The reorganisation would certainly be something to do in the next version's
branch (v7.x-2). We need time to test and improve and enrich and polish
everything. But some feedback about how you think about these steps would be
very welcome!

Comments

rfay’s picture

There is so much in this... Could we have a skype call to talk about it? Would anybody else like to participate? Eaton?

rfay’s picture

The demarcation between amazon API and amazon_media also remained unclear to
me - has it just been an arbitrary decision to make the API save images and to
leave some type-specific fields (for books, music etc.) to amazon_media?

I've never really understood this either. My bet is that Amazon Media was an add-on to the original project, to add more information without destabilizing the existing stuff. But I suspect that it doesn't make any sense in the existing context.

Eaton's goal was to keep amazon.module pure API as you know.

IMO we could refactor in a 7.x-2.x branch.

Also, I would have no objection to reimplementing and integrating features of amazon store into an Amazon submodule. It could use reimplementation, of course, as it has a lot of history, and like all the projects we've done over the years, leaves lots to be desired in retrospect.

rfay’s picture

Issue tags: +amazon-2.x

tagging

shawn dearmond’s picture

subscribing

Dragorth’s picture

This sounds like a useful idea, but will this allow me to possibly integrate my amazon items into a drupal commerces store as an option rather than use just the amazon store. I have a pc repair shop, and would like to be able to add services to my amazon shop, for easy payment purchases. I believe this would require a payment module that redirected to amazon's site for the items not paid for in the cart, and I am not sure how hard a task this would be, but it may be easier that maintaining your own store module, and add the veratility of drupal commerce, which seems to be shaping up fairly well.

Thanks for all the hard work guys, I love the modules so far, and can't wait to see the creative ideas you come up with next!