donebizkit

I used a mix of golang and nodejs. It took a little less than a week to go over the literature and implementation.

Thanks. Once the articles are clustered and sorted, it picks the first news source that reported on the tag (seems fair). The subsequent articles are picked by closeness and similarity.

Bookmarked! Thanks.

Cool. I'll check Asana out.

Nope. I think we are on the same page overall. I just wanted to reassure the new users on this thread that the general practice of submitting a "Show HN" under "Ask HN" is not considered as an exploit. Add to that your point that it should not be abused and I think we have a solid rule of thumb.

Many of us don't have influential networks. We consider HN as the de facto community for hackers and entrepreneurs. Submitting a Show HN is not an exploit. It is a proud move to showcase ones hard work to a fellow audience. I personally think that the community is healthy enough to balance between all types of submissions: articles, Ask HNs and Show HNs in terms of both the people submitting and the people consuming and upvoting. What I agree with you is the need for a separate Shown HN tab similar to Ask HN. There are third party crawlers that do that but we spend all out time here; might as well have the "Show HN" tab here as well.

I submitted my side project last Tuesday @6am and it drowned within 10 min. I found a better exposure when I submitted to ASK HN instead. It would be great if we get the same analysis on the "ask" page. Thanks mayank for sharing your work.

Tx for the hint. The homepage definitely needs some design. As far as the app goes, I purposely want it clutter free so that all the focus is on notes and it's easier to focus on the information in hand (at least as far as I am concerned). What do you think?

I hooked up Ctrl+Shift+M to create new notes.

A good insight on tools and services in the Twitter ecosystem.

I have a long experience in enterprise software and I agree with the premise.

There are two kinds of test units: workflow and functional.

1 - Workflow test units are a waste of time because no single test unit stays valid when there is a change. In other words, whenever we added/removed steps in the workflow, 99% of the time we have to change the test unit to fit that new workflow which breaks the concept of "write once, test all the time" concept. In my experience, having proactive developers who test areas around the workflow that they changed is much faster and reliable.

2 - Functional test units are great. They test one function that needs certain parameters and is expected to spit a certain output i.e function to calculate dollar amounts or do any king of mathematical operations. However, these functions tends to stay unchanged during the lifetime of a project. Therefore, the test units are rarely run.

From my experience workflow changes/bugs represent 80% of the problems we face in enterprise software. Functional changes/bug are rare and can be detected quickly.

This is why I agree with the author premise that unit testing is overhyped.