If you're going to ask users about features, do it over Web and keep it public.
Some people: are deaf or hard-of-hearing; are mute, stutter or are otherwise speech-impaired, or do not speak English; are shy or telephonophobic; sleep, work or are otherwise unavailable during your chosen hours; want to have a persistent record of their feedback and/or the following discussion.
Taking feedback over the phone discriminates against all of the above.
Public service announcement: Building your own dream mechanical keyboard will cost you not much more than you would pay for any decent mass-produced keyboard using the same kind of switches.
Huh? In my version of the reality, RSS is a mechanism by which you subscribe to a source of regular posts such as a blog or a webcomic so that (1) you don't have to visit all these sources' sites every day to see if anything has been posted, and (2) never miss a post.
If you're using RSS as a "here's a bunch of things that went live at these sites you didn't choose yourself and that you might or might not want to read" list... well then I'm sure a Web Extension comes around that can replace Live Bookmarks for you.
Running an RSS retriever on localhost has the same deficiences as using a desktop RSS reader: it’s tied to one device, and it misses posts if turned off for a long period of time.
Whatever you already have. I assume any sufficiently nerd household has multiple computing devices and an ISP that only gives you one public IP address so you have to run a boundary router/NAT box. So I run TT-RSS in an LXC container on that router. And you don’t need a domain name in order to run a server, but if you want that, free-of-charge dynamic DNS providers are a thing.
So run Tiny Tiny RSS on your own server. Or Go Read. Or any other self-hosted RSS daemon. (Yes, these do use accounts and cookies.)
If you rely on Firefox to poll your feeds, you risk missing posts if you are away from your computer for a while, or seeing dupes if you use more than one.
Seriously, just about every HTML mail message I get from my coworkers has small font size, non-default font face, non-black text, a humongous signature, or any combination of the above.
Bonus points for also dealing with discrete contramotion whole-thread-in-reverse-order messages.
A machine with two Ethernet ports can be called a "router" only in a very narrow sense. Pretty much the only kind of routing you can do with those is "192.168.0.0/24 dev eth0; 0.0.0.0/0 via xx.yy.zz.ww dev eth1". (Well, and any VPNs that go over your primary WAN connection.) Your provider goes down, you are stranded on land for the next day or three.
I would like one LAN and at least two independent WAN ports. Three is better. In an inexpensive compact form factor.
Okay, you can stop ranting about lack of vertical space on 16:9 and 21:9 monitors right about now.
You know what I do with my 16:9 monitor? I tile two windows side by side. Boom, instant two 8:9 surfaces, each wide enough to display a terminal 75 lines high and 120 columns wide.
Now, if I could also persuade every web site designer that 960px is a reasonable browser width
Has someone solved the problem of keyboard accessibility of the header bar, in a way that will be consistent across all applications that have it? In a normal application with a normal title bar and menu bar, I know I can press Alt+Space or Alt+underlined letter and access every function. Now, let’s take for example GNOME Calculator. In the top left corner of its header bar, I see a calculator icon that pops up a menu when clicked. How do I access this menu from the keyboard?
The classic title bar performs several functions of varying utility. Let me count them.
1. As the title suggests, the title bar displays the title of the window. This typically includes the name of the application and the name of document currently opened, and can easily take half the space available or even more.
2. It lights up when the window is active, and dims down when inactive, helping the user maintain focus with a busy desktop.
3. It provides an intuitive, discoverable way of dragging the window. (For experienced users, Alt+dragging is more usable, although less discoverable.)
4. It is a big target for (un)maximization via double click.
5. It is a big target for opening the window control menu via right button click.
6. It houses the window manager controls.
7. Last but not the least, the title bar is provided by the window manager in a manner consistent across the desktop. If every application toolkit starts doing its own header bars, we lose this consistency.
I updated on my home machine but not on my workstation.
What *stops* me from upgrading is the lack of a userstyle extension that synchronizes my styles. Not my external style subscriptions — styles that I wrote myself and have no intention to share with anyone. (The technical problem syncing these is that syncable storage is limited to 100KB per extension, and local styles can easily exceed that.)
What I will be *missing sorely* is opening a new tab by entering an URL or a search query in the address bar and pressing just Enter. I will have to train myself to press Alt+Enter instead.
What *saddens* me is that the story for extension UI now seems to be “roll your own in HTML and CSS”. This leads to every extension looking differently, using a different font and different widgets.
I like the new looks of the tabs, though. Praktisch, quadratisch, gut.
Well, that was your problem: Using a text-based container where you are only interested in binary payload. You get all the cost (parsing time and complexity) with none of the benefits (readability).
If that binary is a byte-by-byte representation of some data structure, you could have represented each member of the structure as a JSON object property. No need for binary.
If that binary is just a byte array, you could also use arrays of numbers. Not a very tight encoding, though — 2 to 4 JSON characters per byte of payload.
What you could not do, though, is treat a byte array as a string.
Anyway, what you lacked was not a well-defined specification; JSON is quite well defined. You lacked usage guidelines.
YAML tries to be human-readable and human-writable. Fails on that point. Has as many as five different syntax for string values (double-quoted, single-quoted, unquoted; folded blocks and literal blocks). Even glancing through the YAML spec to find the above information gave me a headache.
Slashdot Mirror
If you're going to ask users about features, do it over Web and keep it public.
Some people: are deaf or hard-of-hearing; are mute, stutter or are otherwise speech-impaired, or do not speak English; are shy or telephonophobic; sleep, work or are otherwise unavailable during your chosen hours; want to have a persistent record of their feedback and/or the following discussion.
Taking feedback over the phone discriminates against all of the above.
Okay, you should probably not game in environments where you get in trouble if it takes a confirmation and 15s for a game to quit.
Well, my second monitor is for one more two-window split.
You're using your widescreen monitor wrong. Tile two windows side-by-side. Read documentation in one half, edit code in the other.
Of course, in that configuration, the only reasonable solution for tab hoarders is (a) grouping into several windows, and (b) multi-row tab bars.
Public service announcement: Building your own dream mechanical keyboard will cost you not much more than you would pay for any decent mass-produced keyboard using the same kind of switches.
You will be baked, and there will be cake.
Huh? In my version of the reality, RSS is a mechanism by which you subscribe to a source of regular posts such as a blog or a webcomic so that (1) you don't have to visit all these sources' sites every day to see if anything has been posted, and (2) never miss a post.
If you're using RSS as a "here's a bunch of things that went live at these sites you didn't choose yourself and that you might or might not want to read" list... well then I'm sure a Web Extension comes around that can replace Live Bookmarks for you.
Running an RSS retriever on localhost has the same deficiences as using a desktop RSS reader: it’s tied to one device, and it misses posts if turned off for a long period of time.
Whatever you already have. I assume any sufficiently nerd household has multiple computing devices and an ISP that only gives you one public IP address so you have to run a boundary router/NAT box. So I run TT-RSS in an LXC container on that router. And you don’t need a domain name in order to run a server, but if you want that, free-of-charge dynamic DNS providers are a thing.
So run Tiny Tiny RSS on your own server. Or Go Read. Or any other self-hosted RSS daemon. (Yes, these do use accounts and cookies.)
If you rely on Firefox to poll your feeds, you risk missing posts if you are away from your computer for a while, or seeing dupes if you use more than one.
Somebody should invent Reader Mode for email.
Seriously, just about every HTML mail message I get from my coworkers has small font size, non-default font face, non-black text, a humongous signature, or any combination of the above.
Bonus points for also dealing with discrete contramotion whole-thread-in-reverse-order messages.
I have seen people do a five-minute voice call to come to a conclusion that could have been reached in half an hour of Slack.
A machine with two Ethernet ports can be called a "router" only in a very narrow sense. Pretty much the only kind of routing you can do with those is "192.168.0.0/24 dev eth0; 0.0.0.0/0 via xx.yy.zz.ww dev eth1". (Well, and any VPNs that go over your primary WAN connection.) Your provider goes down, you are stranded on land for the next day or three.
I would like one LAN and at least two independent WAN ports. Three is better. In an inexpensive compact form factor.
With that kind of tactics against extremists, the Russian government *are* extremists. The defense is making more damage than the threat.
Why the F is the class for the Send button named .T-I-atl, that is the question.
I will welcome any Gmail redesign that will rename classes to readable, understandable English words accurately reflecting their functions.
It would be nice if "Russia" (whatever that means) actually cared about protecting their own people from terrorist attacks.
What they actually care about is protecting themselves and staying in power while labeling people who want privacy as terrorists.
Okay, you can stop ranting about lack of vertical space on 16:9 and 21:9 monitors right about now.
You know what I do with my 16:9 monitor? I tile two windows side by side. Boom, instant two 8:9 surfaces, each wide enough to display a terminal 75 lines high and 120 columns wide.
Now, if I could also persuade every web site designer that 960px is a reasonable browser width
Has someone solved the problem of keyboard accessibility of the header bar, in a way that will be consistent across all applications that have it? In a normal application with a normal title bar and menu bar, I know I can press Alt+Space or Alt+underlined letter and access every function. Now, let’s take for example GNOME Calculator. In the top left corner of its header bar, I see a calculator icon that pops up a menu when clicked. How do I access this menu from the keyboard?
The classic title bar performs several functions of varying utility. Let me count them.
1. As the title suggests, the title bar displays the title of the window. This typically includes the name of the application and the name of document currently opened, and can easily take half the space available or even more.
2. It lights up when the window is active, and dims down when inactive, helping the user maintain focus with a busy desktop.
3. It provides an intuitive, discoverable way of dragging the window. (For experienced users, Alt+dragging is more usable, although less discoverable.)
4. It is a big target for (un)maximization via double click.
5. It is a big target for opening the window control menu via right button click.
6. It houses the window manager controls.
7. Last but not the least, the title bar is provided by the window manager in a manner consistent across the desktop. If every application toolkit starts doing its own header bars, we lose this consistency.
I updated on my home machine but not on my workstation.
What *stops* me from upgrading is the lack of a userstyle extension that synchronizes my styles. Not my external style subscriptions — styles that I wrote myself and have no intention to share with anyone. (The technical problem syncing these is that syncable storage is limited to 100KB per extension, and local styles can easily exceed that.)
What I will be *missing sorely* is opening a new tab by entering an URL or a search query in the address bar and pressing just Enter. I will have to train myself to press Alt+Enter instead.
What *saddens* me is that the story for extension UI now seems to be “roll your own in HTML and CSS”. This leads to every extension looking differently, using a different font and different widgets.
I like the new looks of the tabs, though. Praktisch, quadratisch, gut.
You’ve never used a --force option? That’s the command line equivalent of a checkbox.
Well, that was your problem: Using a text-based container where you are only interested in binary payload. You get all the cost (parsing time and complexity) with none of the benefits (readability).
If that binary is a byte-by-byte representation of some data structure, you could have represented each member of the structure as a JSON object property. No need for binary.
If that binary is just a byte array, you could also use arrays of numbers. Not a very tight encoding, though — 2 to 4 JSON characters per byte of payload.
What you could not do, though, is treat a byte array as a string.
Anyway, what you lacked was not a well-defined specification; JSON is quite well defined. You lacked usage guidelines.
+1 on comments. -1 on YAML.
YAML tries to be human-readable and human-writable. Fails on that point. Has as many as five different syntax for string values (double-quoted, single-quoted, unquoted; folded blocks and literal blocks). Even glancing through the YAML spec to find the above information gave me a headache.
Using a library does not solve the problem of designing a schema, which is what the “property or data” question is all about.
Bullshit. XML *is* part of the Java ecosystem. JSON is from Javascript which is an entirely different thing.