Streams
Streams copied to clipboard
jscheiny
Constructive Feedback
I was told this may be helpful by the party that linked me to the project...
All of this pertains to the documentation of the project
- Link to:
http://jscheiny.github.io/Streams/from theREADME.md - Add some code samples to the above page. I was 3 pages in and had no idea what this was until I got frustrated and clicked over to the code. In general, the Introduction does little to get me excited.
- Would be better to move the API reference to a separate page that is labeled "API reference"
- I have a hard time imagining what the advantage of this library is; please show it off. The basic statistics thing was interesting. I wager that there's something nice in here with respect to either concurrency or the memory footprint; please show that!
- Where are the bloody unit/integration tests? I <3 tests. Not only do they indicate that it works, but they also show me more cool examples.
- Would love some benchmarks compared to the "conventional" way of achieving the same tasks. Not only runtime, also code size and clarity.
All of the above said, this looks to be pretty cool -- well done :+1:. I'll be interested in revisiting this project in a couple of weeks, hoping that its potential awesomeness is advertised better. An impressive bit of work for a personal project.
Sorry for the delayed response, I don't have a lot of bandwidth at the moment for this project. Thanks for the feedback, I will definitely take this into account, and be working to respond to each of the things you mentioned.
Hi Jonah, Great Job! I like the idea behind the library, I was wondering how much work it would take to make this multithreaded? As I had previously started doing this with the STL itself. Interested? As this would be the next natural evolution of your project as it seems feature complete to me.
Hey, Parallelization is something I want to work towards. I'm going to open a new issue to track this.
+1 for Benchmarks. A throughput benchmark that illustrates throughput and end-to-end latency with some decently large file, could certainly add much value in making the point.
Was going through the code to see the 'lazy calculations' part (I am concerned about duplicate computations and unnecessary bufferings etc.) and some pointers on that info will also help.
