marcusegues / newsreader

Single-page full-stack app nspired by Feedly, built on with Ruby on Rails and React.js.
http://swissfeeds.marcusegues.io
1 stars 0 forks source link

Seed Data #5

Open chf2 opened 8 years ago

chf2 commented 8 years ago

I think that this application will definitely benefit from having some well curated feeds as seed data for a guest login, since you may run into the problem of unpredictable data by allowing the demoing user to start directly looking for new feeds. Typically it takes about a day to pull seed data together, so I would encourage you to plan for that around the same time that you are doing styling. If you collect good RSS feed data in the process of working through the parsing and things like that, you can probably cut down on this time at the end.

marcusegues commented 8 years ago

The two essential components are CategoriesIndex and FeedsWindow. FeedOptions is also important because there is a link there to add a new RSS subscription, but I can put this in the CategoriesIndex component at first.

Under this initial simplified version, CategoriesIndex will render list items, each being an "RSS subscription", or "RSS Feed", or "FeedSource. Upon clicking one of the latter, the FeedsWindow will populate with feeds from that subscription.

The way I thought it out, a user would have specific RSS feeds (FeedSources) and the must_read from a FeedSource is true if the user has chosen this FeedSource as part of the Feeds he wishes to be included in a specific Folder/Category called "MustReads". Now that I think of it, "Must Read" can be encompassed by the model FeedCategory, which serves precisely the purpose of defining folders/categories.

I agree that category/folder names should be unique.

Each feed from that is received from an RSS Subscription is a FeedItem. I would gladly store as JSON, but at this stage I have huge questions about what RSS is really and what it gives you and if it is standardized between sites. I'm going to spend a lot of time reading up on this tomorrow, and checking out the repositories you sent.

It is not clear to me either if we should only be fetching the feeds at the frontend as the user requests them (by clicking on a FeedSource in the CategoriesIndex to display in the FeedsWindow), or if we also need to have a certain amount of historical feeds that the user might be interested in stored at the db level. Probably depends on what is the faster way of obtaining old feeds when the user wants them.

I will definitely populate the "Try it Out" version of the site with some reliable feeds.

My initial impressions given your feedback: the site has a simple layout, and the user interaction is basically choosing which subscriptions to display, or adding a new subscription. The most difficult part (on which I am most clueless at this stage) is dealing with RSS. Am I missing something?

On Mon, Dec 14, 2015 at 5:52 PM, Charles Francis notifications@github.com wrote:

I think that this application will definitely benefit from having some well curated feeds as seed data for a guest login, since you may run into the problem of unpredictable data by allowing the demoing user to start directly looking for new feeds. Typically it takes about a day to pull seed data together, so I would encourage you to plan for that around the same time that you are doing styling. If you collect good RSS feed data in the process of working through the parsing and things like that, you can probably cut down on this time at the end.

— Reply to this email directly or view it on GitHub https://github.com/malans/newsreader/issues/5.

marcusegues commented 8 years ago

Just to be clear: "RSS subscription", or "RSS Feed", or "FeedSource" are the same thing.

On Tue, Dec 15, 2015 at 2:52 AM, Marcus Egues marcusegues@gmail.com wrote:

The two essential components are CategoriesIndex and FeedsWindow. FeedOptions is also important because there is a link there to add a new RSS subscription, but I can put this in the CategoriesIndex component at first.

Under this initial simplified version, CategoriesIndex will render list items, each being an "RSS subscription", or "RSS Feed", or "FeedSource. Upon clicking one of the latter, the FeedsWindow will populate with feeds from that subscription.

The way I thought it out, a user would have specific RSS feeds (FeedSources) and the must_read from a FeedSource is true if the user has chosen this FeedSource as part of the Feeds he wishes to be included in a specific Folder/Category called "MustReads". Now that I think of it, "Must Read" can be encompassed by the model FeedCategory, which serves precisely the purpose of defining folders/categories.

I agree that category/folder names should be unique.

Each feed from that is received from an RSS Subscription is a FeedItem. I would gladly store as JSON, but at this stage I have huge questions about what RSS is really and what it gives you and if it is standardized between sites. I'm going to spend a lot of time reading up on this tomorrow, and checking out the repositories you sent.

It is not clear to me either if we should only be fetching the feeds at the frontend as the user requests them (by clicking on a FeedSource in the CategoriesIndex to display in the FeedsWindow), or if we also need to have a certain amount of historical feeds that the user might be interested in stored at the db level. Probably depends on what is the faster way of obtaining old feeds when the user wants them.

I will definitely populate the "Try it Out" version of the site with some reliable feeds.

My initial impressions given your feedback: the site has a simple layout, and the user interaction is basically choosing which subscriptions to display, or adding a new subscription. The most difficult part (on which I am most clueless at this stage) is dealing with RSS. Am I missing something?

On Mon, Dec 14, 2015 at 5:52 PM, Charles Francis <notifications@github.com

wrote:

I think that this application will definitely benefit from having some well curated feeds as seed data for a guest login, since you may run into the problem of unpredictable data by allowing the demoing user to start directly looking for new feeds. Typically it takes about a day to pull seed data together, so I would encourage you to plan for that around the same time that you are doing styling. If you collect good RSS feed data in the process of working through the parsing and things like that, you can probably cut down on this time at the end.

— Reply to this email directly or view it on GitHub https://github.com/malans/newsreader/issues/5.