Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Outputs #1

Open
tlrobinson opened this issue Nov 27, 2012 · 10 comments
Open

Outputs #1

tlrobinson opened this issue Nov 27, 2012 · 10 comments

Comments

@tlrobinson
Copy link
Member

I think there should be an output for each property of an RSS entry (title, description, url, etc), rather than just one called "entries" (also all output names should be singular, right?)

@matthewhudson
Copy link
Member

I think an output for each entry of an RSS works best with the pubsubhubbub RSS Trigger, whereas this is a RSS Action.

Tom - did I understand your suggestion correctly?

Jeff - can you weigh in here?

@tlrobinson
Copy link
Member Author

This isn't really an action, is it? Combined with a "cron" trigger it would be a polling RSS trigger, so I'd expect it to have the same outputs as a pubsubhubub RSS trigger.

@matthewhudson
Copy link
Member

I think it's firmly in the action camp. You could combine it with the cron
trigger, but you would still need a dedup action or some other block to
only display the newest item in the feed.

On Tue, Nov 27, 2012 at 5:14 PM, Tom Robinson notifications@github.comwrote:

This isn't really an action, is it? Combined with a "cron" trigger it
would be a polling RSS trigger, so I'd expect it to have the same outputs
as a pubsubhubub RSS trigger.


Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-10779930.

V/R

Matthew Hudson
Website http://www.matthewghudson.com/ | Bloghttp://blog.matthewghudson.com/
| Facebook http://facebook.com/matthewghudson |
Google+https://plus.google.com/u/0/103902048207396109122/posts|
Twitter http://twitter.com/matthewgh

@progrium
Copy link
Member

Yeah, I think most output names should be singular unless you really are
nesting. And if you're nesting you might not be doing it right because, for
example, this seems like it should return an array of outputs that are
entries. Is that how pubsubhubbub works? I don't see why this would work
any different, otherwise, the pipeline executer won't know what to do with
them ... although maybe that's intentional?

On Tue, Nov 27, 2012 at 2:26 PM, Matthew Hudson notifications@github.comwrote:

I think it's firmly in the action camp. You could combine it with the cron
trigger, but you would still need a dedup action or some other block to
only display the newest item in the feed.

On Tue, Nov 27, 2012 at 5:14 PM, Tom Robinson notifications@github.comwrote:

This isn't really an action, is it? Combined with a "cron" trigger it
would be a polling RSS trigger, so I'd expect it to have the same
outputs
as a pubsubhubub RSS trigger.


Reply to this email directly or view it on GitHub<
https://github.com/webpipes/block-parse-rss/issues/1#issuecomment-10779930>.

V/R

Matthew Hudson
Website http://www.matthewghudson.com/ | Blog<
http://blog.matthewghudson.com/>
| Facebook http://facebook.com/matthewghudson |
Google+https://plus.google.com/u/0/103902048207396109122/posts|
Twitter http://twitter.com/matthewgh


Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-10780404.

Jeff Lindsay
http://progrium.com

@tlrobinson
Copy link
Member Author

Well, if the two options are "trigger" and "action" then it's an action, but I'd rather just call it a block since it has outputs. But that's irrelevant.

The way I see the "dedup" block working is it would have a "key" input used for identifying duplicates, and a "" input you could use to passthrough the rest of the fields, which would be mirrored as outputs (so maybe we need a "" output as well?). You'd connect the individual RSS field outputs from the RSS parse block to the dedup block, then each output to the next block.

Granted this approach will be tedious for records with many fields, but if you just connect an "entry" output then the field names can't be propagated to "downstream" blocks.

Also I'm still a bit confused and concerned about how blocks that output multiple records will work. Say you have two blocks that can output multiple records connected to two different inputs on a 3rd block, how does the executor "merge" the records?

@progrium
Copy link
Member

Also I'm still a bit confused and concerned about how blocks that output
multiple records will work. Say you have two blocks that can output
multiple records connected to two different inputs on a 3rd block, how does
the executor "merge" the records?

In theory it might have a lot to do with the dependency resolution and
which one is tied to a trigger. If they both are maybe it gets weird. But
that's why in theory is hard. Do you have a concrete example that we can
try to make work?

Jeff Lindsay
http://progrium.com

@tlrobinson
Copy link
Member Author

The problem is I can't think of a concrete example, but right now there's
nothing stopping you from setting up such a pipeline in the pipeline editor
/ pipeline definition. I'm wondering if / how we should constrain pipelines
so nonsense pipelines aren't possible.

Maybe restricting a pipeline to having a single trigger or other block that
can output multiple records, or something like that.

On Tue, Nov 27, 2012 at 3:44 PM, Jeff Lindsay notifications@github.comwrote:

Also I'm still a bit confused and concerned about how blocks that output
multiple records will work. Say you have two blocks that can output
multiple records connected to two different inputs on a 3rd block, how
does
the executor "merge" the records?

In theory it might have a lot to do with the dependency resolution and
which one is tied to a trigger. If they both are maybe it gets weird. But
that's why in theory is hard. Do you have a concrete example that we can
try to make work?

Jeff Lindsay
http://progrium.com


Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-10783132.

@progrium
Copy link
Member

Yeah I was already under the impression there's only one pure trigger.

On Tue, Nov 27, 2012 at 3:53 PM, Tom Robinson notifications@github.comwrote:

The problem is I can't think of a concrete example, but right now there's
nothing stopping you from setting up such a pipeline in the pipeline
editor
/ pipeline definition. I'm wondering if / how we should constrain
pipelines
so nonsense pipelines aren't possible.

Maybe restricting a pipeline to having a single trigger or other block
that
can output multiple records, or something like that.

On Tue, Nov 27, 2012 at 3:44 PM, Jeff Lindsay notifications@github.comwrote:

Also I'm still a bit confused and concerned about how blocks that
output
multiple records will work. Say you have two blocks that can output
multiple records connected to two different inputs on a 3rd block, how
does
the executor "merge" the records?

In theory it might have a lot to do with the dependency resolution and
which one is tied to a trigger. If they both are maybe it gets weird.
But
that's why in theory is hard. Do you have a concrete example that we can
try to make work?

Jeff Lindsay
http://progrium.com


Reply to this email directly or view it on GitHub<
https://github.com/webpipes/block-parse-rss/issues/1#issuecomment-10783132>.


Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-10783394.

Jeff Lindsay
http://progrium.com

@tlrobinson
Copy link
Member Author

But allowing any block to output multiple records is also an issue.

@progrium
Copy link
Member

Right, because it effectively becomes a trigger. Hmm..

On Tue, Nov 27, 2012 at 4:04 PM, Tom Robinson notifications@github.comwrote:

But allowing any block to output multiple records is also an issue.


Reply to this email directly or view it on GitHubhttps://github.com//issues/1#issuecomment-10783715.

Jeff Lindsay
http://progrium.com

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants