Carbon – Create images of your source code
github.com
github.com
It would be cool if there was a way to copy paste the text easier.
Like, just using text? Maybe even send it through a syntax highlighter first?
I can't be the only one that was irked at Amazon for spending years making their product promo pages or prominent buttons as hard-coded images instead of text. Horribly inaccessible, over the top, looks terrible on HiDPI devices, looks terrible on a non-Mac because it was clear that their designers were using OS X font rendering, so it looked totally inconsistent.
My favorite part is being on mobile and having to scroll left and right to see the image, or the image has been fit-to-scale is flatly impossibly too small to read.
I feel stodgy, but this entire effort is annoying. Just like the stupid GIFs on project READMEs that play too slow or too fast and have to be watched 50 times to grasp what's going on. Meanwhile asciinema exists and is superior in every way (seekable, pausable, text-selectable, etc).
tarboreus, what's your experience with alt text for this kind of thing?
In a related manner, I'd actually be curious to hear in general about other things we do/don't do that makes life more difficult for blind people on the web.
By my understanding (as a sighted person who has relatively zero experience using such features) alt text is for short pieces of text that might aid in communication of an image to a blind person. I have even less understanding on whether or not dumping a chunk of code in there would be at all useful, and I expect very few sighted persons would know better. Hence my complaint, and hence my question.
"The alt text should be the most concise description possible of the image’s purpose. If anything more than a short phrase or sentence is needed, it would be better to use one of the long description methods discussed in complex images."
The alternatives for long descriptions are the "longdesc" attribute, "figure" with "figcaption" tag or the "aria-describedby".
In short, if the image has alt text, it's fine. But this seems likely to be used in ways that mean uploads without alt text.
Paintings, and Prints, tend to take real world subjects and artistically adapt their native format into a 2d perspective.
This transformation typically removes fidelity oringinally found in the subject. Eg, you can't (or shouldn't) eat a painting of an orange, nor could one typically feel the texture of an orange via a painting.
Alternatively there may be some benefits to the transformation. I'm not sure but in this case one may be able to display more content via the image on Twitter, then transmitted via the oringinal text format due to tweet length "string" hard limits vs image size limitations.
Jesus, at least change the name. Carbon is [already a thing](https://github.com/graphite-project/carbon).
At least, plonk the project name into the github search, and see if there are existing projects using exactly that name. I found at least EIGHT other projects with more then 100 stars already named "carbon". At least one of them is VERY widely used (carbon, which is the time-series backend for graphite).
Basically, you're actively making life harder for users when you decide to name-collide with existing active projects.
Can't help but weep for the excess bandwidth of sharing code as a .png -.-
I know that there probably aren't a huge amount of PHP devs around here, but one of, if not the most, popular PHP libraries in production.
It's not a huge deal I guess, not being in the same space, but I just thought you should know.
Basically, every place you're likely to read code is going to look better than the images generated with this.
See Noel's comment (at the bottom) of https://social.technet.microsoft.com/Forums/en-US/768bd013-c...
Legitimately what is the point?
Three flat-colored circles in a row on a flat surface do not look like realistic items at all, they are actually the opposite, a stylized representation of a real life object.
In this context, the "spirit" is indicating that simulating a "mac window", which serves no major purpose in elucidating source code, which is analogous (but not strictly identical, of course) to simulating physical buttons in another context.
Something like Carbon would have made the illustration a lot better, and only added a few seconds to the process.
Any chance you could add display of that cursor back in?
• It frames the code properly; this is an aesthetic choice, and certainly subjective, but many people seem to prefer it. At the very least, it provides consistent, even margins around the code which are fiddly to achieve with a screenshot.
• If you paste a normal PNG screenshot on Twitter, it gets converted to JPEG and downsampled, creating ugly block artefacts around the text and changing the colours. You can avoid this by making at least one pixel of the PNG partially transparent. Carbon does this automatically (but you can of course do it manually with a screenshot).