Take Better Screenshots | Hacker News Reader

Take Better Screenshots

stilldrinking.org

87 pointspavel_lishin59 comments

> Spot them all: the URL, the problem, the browser, how many tabs are open on that browser, the fact that this person has an unusual number of Facebook tabs open and probably isn’t overly invested in their job, some other applications that are open, that they leave system credentials in text files on their desktop, the time, the operating system, that they have a serious investment in someone named Alex, and whether the wifi is connected.

And that is precisely why I would crop my screenshot to only the relevant part.

The downside of uncropped screenshots is information leakage from e.g. your bookmarks, open tabs, unrelated emails, IM popups etc.

Another commentator mentioned annotated screenshots, and now I'm wondering about the "semantic screenshot"; imagine getting a copy of someone's desktop frozen in time with the ability to re-stack windows, check what's currently available on menus, etc.

I turned off the default Username in menu bar on MacOS so I had one less thing to take off my screenshots.

Tangential to screenshots, but when doing presentations, I've taken to creating a second user on my machine to prevent notifications, etc I'd rather not show.

MacOS seems to turn on Do Not Disturb when plugged into unusual external displays or sharing my screen online, which has always been a big help.

My favorite example is that time Kanye shared a screenshot on Twitter and didn't crop out his Pirate Bay tabs for pirating Deadmau5's music software, Serum.

https://twitter.com/deadmau5/status/704876167157907456?s=20

I see a lot of cropped screenshots on support forums, when really it would be better if the poster took the screenshot of the whole window, and then just highlighted the area they are worried about with a red square (or if you are a StackOverflow user, a freehand red circle ;)). Often the answer lies outside the area the poster thinks is relevant and if one could see it immediately, it would save lots of questions and back and forth.

That said, automatically annotated screenshots would be useful, or even better: if someone invents a screengrab tool that would semantically tag all parts of the image based on the data available so its not just a plain image. I.e. screengrab a GUI window and all the text displayed in that window could be added as metadata like titlebar/button/label text and then it becomes searchable. I guess it would need an API that the developer of the app would have to explicitly implement though for most useful results like debug state information...

Re: annotated screenshots, there once was a module that allowed making snapshots of GTK windows as PDFs: https://github.com/nomeata/gtk-vector-screenshot

Edit: here it is in action, with selectable text: https://www.joachim-breitner.de/various/pdf_screenshot_3.pdf

wow, that's awesome - thanks for sharing! :)

I'm surprised it's not more popular though, but I guess it has remained undiscovered like it was for me until now, or it's simply not as useful to others as I would think. Maybe because not all apps are GTK3 apps, or people don't want pdf's because they can't be inlined on a issue ticket like plain old images can, for example?

> people don't want pdf's because they can't be inlined on a issue ticket like plain old images can

Well, this module can also produce SVGs, which, I guess, can be inlined in some issue trackers. But, aside from being GTK3-only, the main reason this sort of thing didn't become popular is simply due the fact that even JPEG screenshots are just good enough and get the job done. Virtually nobody has any need to search for text in older screenshots; selecting text—well, you can't copy-paste from most of real error windows, so you're going to re-type its content anyways; and nobody is bothered with other people's font sizes, aliasing preferences, JPEG artifacts or whatever.

One Thing where SVG isn't good enough is around fonts - PDF can embed fonts needed, thus produce the "correct" result. SVG probably would pick a different fallback font, which can lead to trouble.

Aside from that: jpeg with compression artifacts works well enough, so that even png, giving better quality, often isn't used by non-technical folks.

And unrelated as an anecdote: The IT hotline at a bank where I once worked asked for screenshots as Word documents (my assumption: "everybody" knows Word and so teaching "press 'Print' on your keyboard and then paste into Word" is relatively reliable teachable via phone ... this went to some level where we as developers got screenshots of sessions containing customer info, where customer details were "hidden" using black bars from Word, which of course could have been removed easily)

I feel like this PDF is a portal to an alternate reality where something something Display PostScript something something something something.

Wow.

I want to click at the UI in the PDF :( it's not working :(

iOS 13 allows for sending screenshots in PDF form.

> This educational blind spot leads Person Q to take a carefully framed picture of the 2 percent of their visual field that signaled a problem and email it to Person X, assuming Person X will immediately grok the necessary portion of Person Q’s life experience that motivated them to send it.

Bravo; I have never seen a more perfectly succinct illustration of the fundamental problem of human existence.

It's a struggle to remember that nobody else is inside your head, but damn, that realization makes everything easier to deal with.

To OP -- in the spirit of the article -- it's an uphill task to separate the content from the witticisms here. IMO it would have been a better read with fewer jokes.

The guy is a writer, I think his goal was to entertain more than to inform.

I agree. I gave up and went straight to HN comments.

To _you_, this person wrote https://www.stilldrinking.org/the-episode-part-1 (and 18 more parts, which I only put down because my body required sleep), https://www.stilldrinking.org/programming-sucks, and everything over at https://www.stilldrinking.org/essays.php.

((Who gets the reference?))

I'm not sure how that is relevant to this essay, which I found almost unreadable.

Ah yes, the famous stilldrinking.org where http://www.stilldrinking.org/programming-sucks comes from :D It's the best essay written about programming, ever. Just one quote:

> This file is Good Code. It has sensible and consistent names for functions and variables. It’s concise. It doesn’t do anything obviously stupid. It has never had to live in the wild, or answer to a sales team.

This piece is so littered with ads (one per paragraph) that it is impossible to read without an ad blocker.

That's, sadly, the Web these days.

Even with adblockers, I see the "advertisement" lines remaining in articles. Often every paragraph or two.

With Javascript disabled, I don't see them at all

Depending on the site, I do.

This LA Times article has two "advertisement" lines. The ads themselves don't appear. JS is disabled.

https://www.latimes.com/california/story/2019-10-31/homes-bu...

With /etc/hosts blocking, I still see them, and I like it. It's like a TV that's gone to ads and then glitches out and the show comes straight back on.

That's what Hulu is like with ublock. It goes to show "ads" but just shows a blank screen, and the countdown goes from 30 seconds to 0 instantly.

More and more often with JavaScript disabled I don't see anything at all!

That's why there's so much blank space between the paragraphs! /s

One of those blank spaces is the white square screenshot mentioned in the article. I assumed it was just an ad taking a while to render the first time I scrolled past it.

http://www.stilldrinking.org/blog_images/suzy/take-better-sc...

Reader mode and adblocker and JS turned off. It was absolutely fine to read, just text and images :)

Why would you even try to read something on the internet without an ad blocker?

Since some people agree to the fundamental "deal" that one receives "free" content for seeing add as well and configure an ad blocker more granular (and maybe the best strategy were that sites with abusive amounts/forms of ads simply lose visitors and attention)

When did I sign that contract?

That response tells that you didn't "sign" that. I only have reasons why some don't or limit use of ad blockers. If you have other values I won't try to convince you.

That's because this article is not about screenshots, it's about the human condition with a dialectical analysis centered on the banal medium of the screenshot vis the existential angst we suffer as human beings now that the floodgates of communication have been busted wide open yet we still have very little of substance to relate

This is a person who truly gets it and knows how to paint a picture with words. Try taking a more visceral approach or listening to Baba Ram Dass in order to truly appreciate the mastery of verse contained in this blog post

My advice is to avoid taking screenshots wherever possible, particularly if you want to share something in a public setting. The're an accessibility nightmare and may make your content unreadable for certain groups of people, screen reader users in particular.

What's the alternative?

Learning to describe your problem.

Specific knowledge aside, I am always astonished that people not being able to coherently read and write is getting worse.

That's something YouTube, Snapchat, Instagram and Facebook don't teach I guess?

The two are not exclusive

Exactly - together they make a wonderful synergy.

I die a little inside every time I get a bug report that is just a couple sentences saying that something didn't work.

I also die a little every time I get a bug report that is just an out-of-context downsampled screenshot of some dashboard, where it has been shrunk so the text is not readable, and there is nothing highlighting what the issue is.

I have to really, really fight back the urge to send back a message with nothing but a gif of the Jerry Maguire "Help me help you" scene

i can do that very well thank you, i just don't have the time budget. i also don't have time to spend on deciphering the problem from a description from even a well-spoken user if a single screenshot would be self-explanatory.

of course it's useful to be able to describe the problem, but most of the time there's no ROI and opportunity cost is prohibitive.

Thing is, accurately describing the problem often is 90% of the solution, and the only reason why you can't solve it yourself and are asking for help is because you don't have the skill, knowledge or experience to see what is wrong there. Similarly, if I'm doing the diagnosis, I don't want your description of what you see there (because that's likely misinformed, misleading, or missing the point), I want the actual evidence.

But twitter and texting do teach it, and those are both very popular. Even reddit is mainstream now. People probably read more now than at any other point in history.

I'm very skeptical that the average reading and writing ability has decreased in the last 50 years

Screenshots for asking help are not the same as screenshots for showing to people. A screenshot is a good way for someone who does not know to send something he does not see himself to someone that will see it. You will capture more than you know in a screenshot, and that's the point. So grab large and you might capture an interesting point that, if you knew was important, you could have communicated as text. But of course be careful you will send more than you think (again it's the point) and try not to include sensitive info as in the last article's screenshot.

> Then there are the well-meaning spirits who discovered the developer console and send these with worried emojis:

All too true. The worst I've ever gotten was an email titled "Problem!" (or similar), the body was several of the grimacing and throwing up emojis, and attached was a screenshot of a (long and ugly) SQL query without the attached results.

When I asked if they could send the results over, or at least a copy-pasteable version of the query, they said they had closed SQL Workbench already along with the query and results...

If a tree crashes to the ground but nobody is there to see it, did the tree actually fall?

In this situation I think there was never a real, valid problem in the first place.

You know what would be helpful? A button in the web browser that the user can click, which creates a full page screenshot and dump of a bunch of helpful data points for debugging (e.g. URL, cookies, history, console logs, network requsets, assets, RAM use, IP address, user agent etc etc.)

Basically a neat way to export a pseudo-stack trace from the application environment with screenshot and all potentially useful info.

Firefox at least does the screenshot part.

https://screenshots.firefox.com/

Or even better, if you can reliably reproduce the problem perhaps record yourself doing so using something like https://www.cockos.com/licecap/

As a dev, the worst offenders are not sending the url and a screenshot with new copy annotated within the image file (and an obligatory arrow to content it replaces).

ShareX provides the best windows-based screenshot in my experience.

https://getsharex.com/downloads/

For me, Greenshot is the most convenient. Not only you can select what area you want, you can edit the screenshot (proverbial red circle, text, etc.) before saving it.

https://getgreenshot.org/

Checking out the browsers / desktops of my coworkers when they are sharing screens is fun. We still make fun of one guy who had google bookmarked in chrome.

No, I favor full screen screenshots because people tend to not include important things when they crop.

Also you get to spy on people a little

https://github.com/jordwest/news-feed-eradicator

How is this related?

"Actually, the end of the Facebook feed would be great"

Reply on news.ycombinator.com