<ocdtrekkie>
I'm having not a lot of fun figuring out how I plug in stuff to the convert command directly.
<dwrensha>
so you could try something like `convert -background none -scale 50x50 -negate icons/share.svg share.png`
<dwrensha>
i guess it figures out the formats you want just by looking at the filename extensions
<ocdtrekkie>
Yeah, I figured that.
<ocdtrekkie>
So, the share button is blue so it stands out from the other buttons?
<dwrensha>
right. We don't want people to miss it.
<ocdtrekkie>
Okay. And does your CSS do anything to resize or reorient the icon? (Aka, is my PR going to work right?)
<ocdtrekkie>
In the grand school of things, I think all icons' padding in the image itself should be the same/similar so that unique code isn't needed for each button to make them appear similar in size.
<dwrensha>
the CSS I added doesn't resize or reorient.
<ocdtrekkie>
Okay, so my PR should still do what it's supposed to (in theory). And now I get why the share button gets special CSS.
<ocdtrekkie>
:D
<ocdtrekkie>
At some point, thenounproject started resizing the SVGs for free downloads, I think, in order to accomodate/add credit text to the SVG file.
<ocdtrekkie>
Or it might've just been specific SVGs, I don't know.
<dwrensha>
incidentally, we need a better way to share one-off images with Sandstorm
<ocdtrekkie>
Ah, okay, that is very helpful.
<ocdtrekkie>
dwrensha: My noted suggestion is to have a lightweight ability to just "upload files".
<ocdtrekkie>
Whether or not you have a Sandstorm app that can display them. (Similar to other online storage solutions.)
<ocdtrekkie>
For complex things, it makes a lot of sense to have data inside apps, but in the case of one-off files like images, plain text files, etc. it may be worthwhile to allow people to upload the file, and then choose to open it with apps or viewers that know how.
<ocdtrekkie>
Re: the screenshot itself, I am not sure rounded a rounded corners button looks ideal with the UI, as the current style is entirely squared off. (See the Sandstorm logo, the app title, the highlighting when you mouseover buttons.)
<ocdtrekkie>
I'm inclined to suggest the Share link and blue background should be the full height of the Sandstorm bar.
<ocdtrekkie>
(Arguably, existing share buttons' mousovers should also be full height of the bar.)
bb010g has quit [Quit: Connection closed for inactivity]
<dwrensha>
I just made the mistake of pressing tab while editing a GitHub comment and expecting it to mean "indent".
paulfitz has joined #sandstorm
jadewang has joined #sandstorm
paroneayea has quit [Read error: Connection reset by peer]
jadewang has quit [Ping timeout: 252 seconds]
paroneayea has joined #sandstorm
paulfitz has quit [Ping timeout: 240 seconds]
<paulproteus>
For those who like upvoting things: https://news.ycombinator.com/shownew , look for 'Show HN: “Try a demo” on Etherpad front page (etherpad.org)'
erikoeurch has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 244 seconds]
jadewang has joined #sandstorm
<dwrensha>
I am struggling to understand how the demo server got a big blue share button with the old share icon
<kentonv>
dwrensha: The magic of caching.
<dwrensha>
it's even properly transparent
<kentonv>
it's IN YOUR BROWSER
<kentonv>
:)
<dwrensha>
so it must have been built from the old svg
<kentonv>
meteor names the javascript and CSS as a hash of the content
<dwrensha>
i would understand if it were the old icon an not properly transparent
<kentonv>
so they won't ever hit cache
<kentonv>
but images are named whatever you name them
<kentonv>
so they'll stick around in cache for a while
<paulproteus>
Blargle we should do something smarter for URL generation to avoid this kind of cache insanity.
<kentonv>
I think this "-alpha shape" was an attempt to make the background transparent, and I'm not sure why I didn't just use "-background none" for that, but there are so many options and maybe I just missed it.
<paulproteus>
aldeka: I successfully said hi to zarvox
<aldeka>
yay
<paulproteus>
I told him if he wants to say hi to you back, he can do it himself
<paulproteus>
: P
fonfon has quit [Remote host closed the connection]
<ocdtrekkie>
One of those stupid things that can only come from something as mind-addled as CSS... When you click the Share, the word Share moves left a pixel for like a second on my Firefox.
<ocdtrekkie>
The icon and the button, on the other hand, does not move.
<ocdtrekkie>
I assume it's unintended, as I have no idea how or why that would happen.
natea has joined #sandstorm
<XgF>
The fact that the share icon isn't center aligned in the titlebar is Really Irritating (TM)
<XgF>
(Also, any particular reason for not grabbing all of the icons from a single consistent icon set?)
<kentonv>
XgF: The icon set we were using originally didn't have a good share button.
<ocdtrekkie>
Is the new guy on IRC yet?
<kentonv>
ocdtrekkie: he's zarvox
<paulproteus>
ocdtrekkie: omg he already was
<paulproteus>
: D
<ocdtrekkie>
\o/
<ocdtrekkie>
zarvox: \o/
ragesoss has quit [Ping timeout: 240 seconds]
<ocdtrekkie>
I would think at some point, XgF, Sandstorm would have it's own distinct icon set, at least partially customized from other icons.
ragesoss has joined #sandstorm
<ocdtrekkie>
I've stretched and sized a few to not look odd at a glance, but eventually some manner of special attention to icons would be warranted.
<ocdtrekkie>
(Right now it's just the only tedium I have to occupy myself with since I never finished the Meteor tutorial.)
<XgF>
If it were me I'd just go and grab the Android icon pack :P
<XgF>
They're reasonably bland, have the right share icon, etc
<kentonv>
XgF: We'll eventually replace everything with nena-designed art
<XgF>
:-)
<zarvox>
ocdtrekkie: \o/
<ocdtrekkie>
XgF: If one more site I use daily adopts Google icons/styles/themes/etc. I am going to stab myself in the eye with a spork.
<ocdtrekkie>
And I have two titanium sporks at home.
<zarvox>
quick dwrensha, change the share icon to not be google-blue! :P
<ocdtrekkie>
D:
<ocdtrekkie>
Shouldn't the share icon be :purple:?
<ocdtrekkie>
I was going to ask that earlier but forgot to.
<XgF>
Eh, going full on material design would be one thing, but the icons are pretty nonspecific
<ocdtrekkie>
Cuz the icons are kinda that sandy tan thing. And then the other Sandstorm-ish color now is purple.
<ocdtrekkie>
zarvox: What's the difference between google-blue and cornflower-blue?
<ocdtrekkie>
XgF: The problem with using any foreign icon set is that your software will probably eventually need an icon they don't have. (The first iteration of the menu on Sandstorm only needed buttons that one designer had made.)
<ocdtrekkie>
So then, you end up needing someone to try to emulate the style of the original set. Rather than having one artist create all of it.
<ocdtrekkie>
I do hope future icons are SVGs though. Everything should be SVGs.
<XgF>
ocdtrekkie: Sure. If you have an icon designer, then that's awesome. If you don't, a big one forestalls the issue for a while
<ocdtrekkie>
Yes, but so does searching thenounproject for an applicable SVG file like they're doing now. :)
<phildini>
zarvox: \o/
<zarvox>
ocdtrekkie: I don't know actually; the share button reminded me of the Google Docs share button, and you seemed out for blood ;)
<zarvox>
phildini: \o/
<ocdtrekkie>
zarvox: lol. It should probably be purple though. But it seems cornflower-y, which is an acceptable blue for all tech things.
<ocdtrekkie>
If you add red and yellow though, I will come for you.
<ocdtrekkie>
:P
ragesoss has quit [Quit: No Ping reply in 180 seconds.]
ragesoss has joined #sandstorm
<kentonv>
... what if I said I like material design?
<ocdtrekkie>
There's a lot of people who like it. (I do not understand why. But such people seem to exist.)
<kentonv>
ocdtrekkie: remind me what exactly you don't like
ragesoss has quit [Quit: No Ping reply in 180 seconds.]
<ocdtrekkie>
Biggest issue is Google's apparent belief padding is more useful than content, and that padding is a viable use of mobile screen space, considering mobile screen space eats battery life.
ragesoss has joined #sandstorm
<ocdtrekkie>
Everything about Material Design seems to be intended to make my 5.5" phone screen about as useless as my old 3.7" phone screen.
<kentonv>
so if there is a "compact mode" option, does that make it OK?
<ocdtrekkie>
I'm not fond of the rampant early iPhone skeuomorphism of Material, nor the crayola color scheme, but that's all cosmetic and I can live with it.
<ocdtrekkie>
kentonv: Absolutely.
<kentonv>
cool, I totally agree that should always be an option.
<ocdtrekkie>
IMHO, display density in Gmail, Calendar, etc. was one of Google's best design decisions.
<ocdtrekkie>
Which is why it makes me so sad that every single UI update these days involves removing said options.
erikoeurch has quit [Ping timeout: 250 seconds]
<ocdtrekkie>
In the case of Sandstorm, some optional UI settings would probably be worth passing as API parameters, so apps which want to support it can potentially configure themselves for users automatically.
<kentonv>
I do see how for a lot of people, information density can be a turn-off
<XgF>
Nobody does UI options because they're a massive ongoing pain in the ass
<ocdtrekkie>
For instance I prefer dark themed Android apps because they do not blind me at night and they use a lot less battery on Super AMOLED screens.
<ocdtrekkie>
But most apps that support them I still have to individually configure their UI. I'd love to tell the whole phone "dark backgrounds, darnit".
<kentonv>
XgF: In general I prefer to avoid options, but there seem to be some people who feel very, very strongly about information density in opposite directions
<ocdtrekkie>
XgF: The problem there seems to be that developers do not place adequate value in user choice.
<ocdtrekkie>
Which is funny coming from companies which espouse making computing more "personal".
<ocdtrekkie>
If Google wants to claim their services are personal, why can't I make the background of their mail app black? Google even has a Material Dark theme built right into their operating system's standard themes.
<kentonv>
ocdtrekkie: you do have to balance the value of choice against the cost of implementing choices when the developers could have been working on novel functionality or bugfixes.
<XgF>
ocdtrekkie: because J Average User accidentally turns it on/their child fiddles with the settings and turns it on/etc/etc and now they can't read anything in GMail and they don't understand why and Google sucks and they mess up all their apps and make them ugly
<ocdtrekkie>
kentonv: As long as Google seems busy making things like hamburger menus that twirl around into a back arrow, I am convinced Google's designers literally nothing worthwhile to do on their plate, and should be letting users have important things like display density and dark themes.
<XgF>
But the hamburger twirling around into an arrow is my favorite thing :P
<ocdtrekkie>
XgF: I lost all hope for Google when I saw that. Shortly followed by seeing people's fanboying posts about it, which made me lose all hope for those people. :/
<kentonv>
I absolutely agree that Google does a poor job of prioritizing what they work on
<ocdtrekkie>
Density and color are both huge usability concerns. There's some good articles on contrast issues with Google apps in material themes.
<ocdtrekkie>
In several examples where Google itself ignores Material's readability guidelines.
<ocdtrekkie>
This week, my big accomplishment was moving all of my mobile devices over to a forked version of K-9 Mail, before they go material.
<kentonv>
yeah designers love low-contrast because then all the ugly words don't mess up their pretty design so much (and honestly, if you aren't actually trying to _read_ the text, it does look nicer)
<kentonv>
but it's a terrible idea
<aldeka>
zarvox: \o/
<zarvox>
aldeka: \o/
<ocdtrekkie>
On any given Google page, there's like six or eight different flavors of grey.
<XgF>
kentonv: Apple's uberskinny Helvetica is the best example of this :)
<ocdtrekkie>
And kentonv: Also, if the shell ever becomes an app in itself, as you've indicated you want to eventually do, much design issues becomes less relevant because people can have their own flavors.
<kentonv>
XgF: I sometimes feel bad about using open sans "light" in our web design but if I use anything thicker then it all looks so _dated_. :/
<ocdtrekkie>
I get the idea of putting design > content in a front page that's gotta sell something.
<ocdtrekkie>
IMHO the issue is when you get to the stuff you're supposed to use daily.
<kentonv>
fair enough
<kentonv>
gtg to a meeting
<XgF>
And this is where we'll disagree; IMO the GMail android app is just about right, and I say that as someone who has to sort through an incredible volume of E-Mail
<ocdtrekkie>
I just don't want my email app to assume being pretty is more important than helping me get through the last 200 emails I received as quickly as possible with one hand while I'm also navigating a parking lot.
<ocdtrekkie>
XgF: The one-handed requirement kills Gmail. Google removed the best UI element ever in Gmail, and it's the main reason I love K-9.
<XgF>
?
<ocdtrekkie>
K-9 has (and Gmail used to) the action items on the bottom of the screen.
<ocdtrekkie>
So I can easily reach all of the options I need to sort, read, and delete email with my thumb while my phone is in the same hand.
<XgF>
If I'm sorting by hand, something's gone wrong with my filters. People delete E-Mail? And what options do you need to read?
<ocdtrekkie>
XgF: I essentially use a Gmail folder for browsing social networks. So I get a LOT routing through my mail app. And no, I don't need the entire history of G+ in my inbox.
<ocdtrekkie>
So yes, I delete a LOT of mail.
<ocdtrekkie>
But it's far faster, easier, and more convenient than actually browsing G+ on G+.
<XgF>
Do I need it in my inbox? Probably not. But why would I delete it?
<ocdtrekkie>
XgF: Less useless search results when I need to search my archive.
<ocdtrekkie>
I use my email pretty heavily as a notification tray. So if I didn't delete stuff, I could get 20,000 results for a single search term, many of which would be identical.
<ocdtrekkie>
And "Delete" is as easy to press as "Archive" is.
<ocdtrekkie>
I probably delete 19 emails for every one I archive.
<ocdtrekkie>
This is much the same with my work email as well.
<ocdtrekkie>
Though my work email has an additional 50 spam messages for every 19 I delete and 1 I archive. ;)
<XgF>
I thought about deleting a pile of the automated repetitive daily E-Mails I receive at work. Then searching them actually came in handy one day...
<ocdtrekkie>
That's a great example of why different people may want different things, and email programs should never try to force you to do one thing or the other. ;)
<ocdtrekkie>
See the outcry Google got when it tried to hide/remove the delete button on mobile Gmail.
<ocdtrekkie>
kentonv: That comment thread. o_o
tantalum has quit [Quit: Leaving]
natea has quit [Quit: natea]
<phildini>
does anyone here have opinions or ideas about a self-hosted alternative to new relic/ datadog?
mattl has joined #sandstorm
<XgF>
My only opinion on New Relic is that, gosh, they seem to spend a lot of money advertising to me on twitter :p
<phildini>
lol.
* paulproteus
waves.
<paulproteus>
Hey mattl.
<mattl>
hey paulproteus
<paulproteus>
BTW meet zarvox, the person I mentioned was working on vagrant-spk. (-:
soulshake has left #sandstorm [#sandstorm]
<paulproteus>
I aim to write some glossier (aka more easily-consumable) docs on packaging PHP apps with vagrant-spk next week. If you are excited right now to make a package, though, I can also help you through whatever problems you're having.
<paulproteus>
So let me know what makes sense for your available time, etc.
<paulproteus>
2. I wonder if there's any way to integrate people's self-hosting scrobbling with the music collection you have on libre.fm.
<paulproteus>
Maybe yes, maybe no. But maybe that's not important.
<mattl>
so, GNU FM is federated
<mattl>
we even have support for federating to Last.fm :)
<mattl>
so you can listen to a song on Libre.fm and have it recorded back on your personal GNU FM site in addition to your Libre.fm profile.
<mattl>
paulproteus: you'll have a tough time upgrading all the Audioscrobbler clients in the world, so maybe supporting scrobbling will be a lot easier.
<paulproteus>
I think that seems right, yeah.
<paulproteus>
One principle of Sandstorm is that we want to be able to separate browser-initiated HTTP requests vs. non-browser API clients, to avoid cross-site request forgery-type problems.
<paulproteus>
I'll have to look at the scrobbling protocol to see what it's like.
<XgF>
I think fundamentally theres kind of a need to accept that you aren't going to redesign every protocol in the world to use OAuth2 bearer tokens..
<mattl>
paulproteus: look at zomg, a very simple client written in zsh by Clint.
<paulproteus>
mattl: < 3 Client
<paulproteus>
I mean < 3 Clint
<paulproteus>
(that was a typo, not some smart pun thing)
<XgF>
On othe other hand, if you put the API at api.mysandstorm/random-unguessable-string/, then I'm not sure XSRF should be an issue
<paulproteus>
whatever, we should be able to patch all of those
<paulproteus>
: P
<paulproteus>
(kidding!)
<mattl>
let me know when you've patched iTunes, it's the only client I use.
<paulproteus>
kentonv will write a dtrace module that hotpatches it this weekend probably
<XgF>
Oh look, "Anomaly (Chrome extension)" :D
<paulproteus>
XgF: I am literally laughing out loud.
<mattl>
there's also an unreleased Chrome extension for Google Music, and others.
<paulproteus>
Anyway I'm sure kentonv has been meaning to look into dtrace for 10 years or so, and this would make a great starter project.
<kentonv>
I forget what dtrace is
<mattl>
it's a way of making races out of the dt, command, whatever that is
<paulproteus>
dx/dt
<mattl>
dt is a generic data test program used to verify proper operation of peripherals, file systems, device drivers, or any data stream supported by the operating system.
<mattl>
anyway, yeah. i want "try this" buttons on gnu.io for GNU FM and GNU social, so let's work on that.
<paulproteus>
Cool (-:
<XgF>
GNU Social sounds fun, what with the big public APIs :-)
<mattl>
paulproteus: do you think we could get a special version of https://sandstorm.io/preorder.html too where it says GNU/Linux, so we can link to it for "host your own instance"? I know you know the difference, but the person who writes a cranky email to RMS about it doesn't know that.
* XgF
is tempted to install Sandstorm on a non-GNU Linux install to make some sort of point :P
<paulproteus>
I think https://sandstorm.io/preorder.html?gnu would be fine. I would merge a patch that looks for ?gnu in the querystring and does the text change you mentioned; I would also accept a bug filed instead of that, and might work on it but make no promises about timeline.
<mattl>
hmm, static page.
<paulproteus>
It's fine; Javascript on the page can do the change. It can be an inline script and/or comply with LibreJS.
<mattl>
yeah, trying to think about how best to do it. <noscript>free software</noscript> all over the place ;)
<mattl>
the alternative would just to be have a preorder-gnu.html and it says "Host your favorite GNU applications, including GNU FM, MediaGoblin and GNU social"
<paulproteus>
FWIW for self-installed Sandstorm, you don't need the pre-order page involved.
<mattl>
oh yeah, but I want people to be able to quickly host their own instances of these things.
<mattl>
and really the people who can self-host already do.
<mattl>
anyway, that's my problem not yours.
<mattl>
XgF: i'd be interested in seeing how that goes, for reals.
<XgF>
Non-GNU Linux? TBH, for what Sandstorm needs, you could probably get very close to there with Linux + Musl libc + libc++ + busybox
<XgF>
Probably need to fix up a glibcism or two
<paulproteus>
I like the idea of making the web APIs for GNU social and GNU.fm work, but need to fix some installer bugs first. Until the web APIs are compatible, though, Sandstorm isn't going to be a good fit, mattl. )-: