williamsharkey

To be more explicit, it is insane that we find it acceptable for text to reflow/ break over lines at different points once the web font is loaded. There is something that bugs me -- On the one hand we are being considerate and trying to make the reading experience pleasant. On the other hand we are being careless creators of jank. Maybe jank is something that primarily irritates nerds and I am overreacting.

I wish that system fonts could be told the size of the characters of a web font upfront, before the web font was loaded. The system font's glyphs should smoothly fade/morph into the target webfont once it is loaded.

To solve the problem generally would be difficult or equivalent to the effort/data required to download the web font in the first place, but there may be a quick and dirty solution that covers 99% of text in the wild. The easiest case is monos paced fonts. Is there really an excuse why the system font can't be told via Css what spacing/height to use to match the target webfont?

The author makes some interesting comments, view the page source.

I wish yCombinator or equivalent would fund a company or organization that is dedicated to making a complete end-to-end opensource fpga ecosystem. The proprietary tools are crap.

I would be interested in comparing the performance of Uber's first pass "city-fencing" vs a pre-computed 1D array.

EG: Subdivide a flat projection of earth into n^2 squares. Create an array of length n^2. Set the value of each element in the array to a list of canidate geofences(which have area in that square).

Scale lat and long between 0 and 1. Then you can index directly into it with PrecomputedArray[floor(lat*n+long)]

This is trading space for time, may as well choose space here.

Can you measure the blur that I see when I do not wear glasses, and apply the inverse function to an image, like my computer monitor, so that I can see clearly without glasses? What sacrifices would be made -- dynamic range?

Can the progressive data obscure previous portions of the image?

If so, it can also be a hacky video format.

EG: A flif would specify how many megabytes it used per second. There would not be discrete "frames".

If a video flif said that it used 1MB per Second, and you wanted to see what the video looked like at 10.8 seconds, you would download 10.8MB .

The author writes:

"Sometimes (unless I am writing in Go) I don't want to bundle all my code together into one giant hulking binary."

I am unfamiliar with Go - can someone please offer why this technique might especially desirable/feasible with Go?

Can you fuzz an uninstrumented executable to add instrumentation?

Challenge - Upload a photo of the print, see it in frame.

Extra credit - Take some photos of your frames in natural room settings. Inside the frame, place a color calibration sheet. Find the transformation of the color calibration sheet (EG - how the room affects color, glare, and shadows), and then apply it to the user-uploaded images. Perhaps this could sell the difference between normal and anti glare.

Does Minuum use the Y position of taps as well as the X? If not, it is throwing a predictor away.

Bad algorithms are bad and misused feedback is misused and competent advisers are competent.

I collect contact info from individuals who want a wake up call at the same time and pair them by geographic region so they can phone each other.

This service not automated, complicated or scalable. It uses a Google form to collect information, which is appended to a Google spreadsheet that I have access to. It is hosted on neocities.org, in case I wish to add informational pages in addition to the sign-up form. I manually verify, pair and contact participants.

My goals are modest, 100-300 users, so that the chances of finding people who match are good. If there were more than 300 users I would not be able to administer it manually.

Are there any changes I could make to speed adoption or increase usefulness?