Ask HN: What interesting problems are you working on?

Am working on a travel site focused on showing places with activities and having detailed information from how to get there to pricing since most of information on google is outdated. I had stopped working on it after reading some threads here on why disrupting travel is very difficult but seems easy to a texh person looking for ideas and I almost shelved it. Then I remembered most commenters on here are from western countries that are probably reaching a saturation point of ideas in that almost any app imaginable already exists. Anyway, the renewed interest has made me start from scratch trying to collect the data. Anyone who would be interested in joining me and be the coder while i focus on the rest, i’d be glad to team up. Focus is on building an extensive trip planner which you can view a location and add to “cart” and at the end see the total coast of the trip. This can be helpful a lot for solo travelers. East Africa is still a popular destination. Reach me on ninaformke at g mail

I go on long walks, this is something I will try. I usually listen to podcasts while walking, maybe it is time to try something different.

About analyzing thoughts - I tend to over analyze everything. How do you not fall into this trap?

I've had a few discussions with the author of that article. Not intending to put words in his mouth, but I believe he is effectively saying 2 things:

- The metrics Circadian Stimulus and Melanopic Lux are flawed

- We don't know enough to produce products yet

I agree with him on point 1, especially about CS. The author, and several others, have demonstrated a huge discontinuity around 3500K. The CS model needs improvement.

Regarding melanopic lux, the author says it is more a problem with our measurement and modeling tools. Most manufacturers won't provide a Spectral Power Distribution, but this is starting to change (we do). There is at least one clinical trial in progress that is testing this metric directly, and several experts think it shows a lot of promise.

With that said, new research on blue vs. yellow/orange opponency is coming out. Basically, we're trying to figure out contribution of the visual system on top of melanopsin. I'm not sure what to make of it yet, but I think we'll find out much more in the next 2 years. We probably need a better metric than melanopic lux.

On point 2, I agree that our knowledge is changing. But it takes a number of years for standards to be developed, much slower than the rate of technology improvement. We are paying attention to all the new research, but we based our design on our best understanding of the research to date.

We know that we can really only attract early adopters at this stage who "get" it, and that most people are going to wait and see until things are more standardized. But we are effectively able to do mini-experiments with our customers, which can lead to more insights/avenues of experimentation.

Case-in-point: f.lux. Research has shown that it is somewhat effective on its own, but combined with dimming, it could be quite effective. That's actually what we're doing with our product—controlling both the spectrum and intensity. And f.lux is able to run experiments with a portion of their users to advance our understanding.

Basically, my opinion is that we have to start somewhere, and we believe we have enough evidence supporting that our solution is "better than nothing." We were able to prove that there is enough of a market for these types of solutions already, even though our knowledge is still evolving.

(edited for a paragraph spacing typo and a 2nd time for a grammar issue)

Flight search. I hunt deals for Aussie travellers (https://www.beatthatflight.com.au), but an inordinate amount of my searching for deals is manual.

- initially, all manual

- secondly, timers - I know when some airlines do deals, so I go look

- thirdly, I found other sites indexing unusually cheap flights, but they're not always the same price on my site

- fourth, built a script to search my own site for a route, but the number of combinations rockets with the increase in date ranges. If you're taking different stopovers etc, it becomes ludicrous.

- it's growing at least, but finding ways to make it less hands on and less mind-numbing is a never ending quest. Although I still enjoy it :)

Focusmate does that. Only worked for like 3 Sessions until novelty was gone