Show HN: Open-core icon sets that took me 12 years to create (app.streamlinehq.com)
Hi HN,
Vincent here, maker of Streamline 4.0. I spent the past 12 years perfecting icons and making the largest sets on the internet.
The 5 open-source sets are:
1. User Interface Icon Set (1,000 icons)
2. Streamline Flex (500 icons)
3. Streamline Flat (500 icons)
4. Covid Icons (147 icons)
5. Nasty Icons (45 icons)
134 comments
[ 3.5 ms ] story [ 209 ms ] threadThanks for creating this.
Also it is not clear to me what is the license, if I create a product for a customer and I stop the subscription I guess this product can still use the icons I put them there.
Anyway, congratulations for the product, I bookmark it.
2. Thanks for letting me know that the licensing isn't as clear. You're absolutely right in assuming that your customers can continue using the icons even if you cancel the subscription.
https://creativecommons.org/licenses/by/4.0/
I'll take a look what are the other 6 licenses and make our licensing information clearer. Thank you. I appreicate this.
https://creativecommons.org/licenses/by/4.0/
Other premium icon sets aren't open-source, so using them would require a backlink or a purchase.
CC BY is recognized by the FSF as a free license and by Debian as DFSG-free. While it's not listed by the OSI as open-source, I think that's mostly because it's not a code license. So I don't think that calling it a open source license is misleading.
Of course, for open source icons I would actually expect the source files to be open...
Although, I'd do it differently since I didn't work for 24 hours a day. I'll also account for our open-source illustration and emoji sets.
So, (12 years * 12 months * 4 weeks * 40 hour work week )/(104600 icons + 2322 emoji + 24863 illustrations).
Streamline now has a team of 15 people so I'm not the only one doing it all by myself :)
I also found the licensing confusing. You say it's under CC 4.0 but don't specify. When digging deeper it appears you mean CC BY 4.0, but then you link to another page that places specific restrictions on the type of attribution, which CC BY does not do.
What's more, some of these restrictions are onerous, e.g., requiring attribution in an image's text description or requiring a link to your page. Again, that is inconsistent with the license you claim to use, and what's more is you don't have a policy posted on referrer tracking, so for all we know you're essentially requiring telemetry.
In my estimation, you won't get a lot of people using these until you clear up the licensing. While your icons are beautiful (again, great work!) it isn't worth the legal risk.
https://creativecommons.org/licenses/by/4.0/
Other premium icon sets aren't open-source, so using them would require a backlink or a purchase.
Except for wide practice? How many people who work with mobile phones cannot really determine what the icon for Wi-Fi means?
"So, just suggesting that the Roman alphabet was a step forward from icons and returning to icons is a step backwards."
Backward and forward with respect to what? Writing language down in an alphabet is more precise, but only other speakers of the same languge will be able to communicate with you. Symbolic writing is less precise, but more intelligible to people who do not speak your language.
Would China hold together if they didn't have a common symbolic writing system? The languages (or dialects) spoken in the country are very different from one another.
But that's the problem. You don't know what the icon for Wi-Fi means, until you do.
That is quite a normal task in human civilization, we do not come out of the womb pre-programmed to be literate.
In marketing the goal is to, every step of the way, not lose customers. Basically: You want to sell stuff or attempt to teach stuff? If a simple hamburger already is < 100% everything else must be terrible.
More than 0% wont be able to see the little house as a link to the front page while something like "home" would work for an English audience 100% of the time.
Right, for an icon, commonly, hovering the mouse over the icon causes the display of a word or two or so of English that defines the icon. So, I'm suggesting, just use the word or two of English instead of the icon.
In the case of Unicode, you can use the code name
> 2. type them in text,
In the case of Unicode, you have the code point
> 3. send them in text email,
Same as #2
> 4. spell them,
Same as #1 (though why you'd want to spell a glyph is beyond me)
> 5. sort them,
Same as #2
> 6. look them up in a dictionary
Databases for Unicode and icons already exist
> 7. search on them.
Same as #6
> 8. Have no reasonable and reliable way to determine what they mean.
Here is the clincher for why icons have become so widespread: they're actually easier for international audiences to determine what they mean because they illustrate an action rather than describing that action in a specific language.
They're also easier to parse for anyone with reading difficulties like dyslexia.
There's a reason traffic signs employ iconography. It's actually easier for the lowest common denominator to read and understand. Plus they have a fantastic benefit of conveying information concisely. Another example of that latter point if the iconography on media players (the pause, play, et al buttons).
> At my Web site, I have no icons and, instead, for links use words typed using the Roman alphabet, words in English, words that have meanings in standard dictionaries, can be spelled, typed, etc.
As with most things, it depends largely on your audience/use cases. Your site might not benefit from them but that doesn't mean there isn't a benefit.
> So, just suggesting that the Roman alphabet was a step forward from icons and returning to icons is a step backwards.
People who cannot read English might disagree with you there ;)
For people who don't know English, between an undefined icon and an English word they don't know, which is easier to decode, get defined? The English word can be defined in a dictionary and, then, readily translated to nearly any language on the planet. Nothing similar is available for icons.
I saw. But the point you made was far more generalised so I gave examples of where icons can have that feature.
> The English word can be defined in a dictionary
Which is still no use if you cannot read English
> and, then, readily translated to nearly any language on the planet.
Automated translations are usually terrible. They’re just about passable for larger volumes of text because you can figure out what the article is generally about. But for short words without context, good luck getting a meaningful translation
> Nothing similar is available for icons.
You keep saying that and you keep overlooking that a good icon will illustrate the action in a meaningful way.
Sure shit icons do exist but then so do shit labels for actions that don’t describe the action well. But you cannot form an argument where you’re comparing the worst of icons against the best of written labels. That’s simply not a fair basis for an argument.
You’ve also overlooked my point about conveying information clearly but tersely. And about people with other kinds of reading difficulties. In fact you seem only focused on web site design while making a sweeping statement that all icons are bad.
Honestly the best thing you can do here is go read a book on design. Or on why road signs are the way they are. Or even just an article on why the media glyphs were invented. You’re quoting your own opinion but here but there’s been decades of research into this very topic.
Then my context was also Web pages and not all uses of icons. I certainly was not talking about highway road signs.
> But for short words without context, good luck getting a meaningful translation.
For English words, the meanings are in dictionaries and easily elsewhere online, e.g., Google.
E.g., when I was reading some old math in some old German, I just used Cassell's dictionary between English and German. So, I would take some one word in German, go to the dictionary, and get the definition in English. So, my input to the dictionary was just one word, not a lot of context. Worked fine. So, I was able to read the book on differential geometry by Václav Hlavatý (I was taking a reading course from him -- he had helped Al) translated from his Czech to some pompous old German -- I found the old German difficult reading and so did a fellow math grad student from Germany. Still Cassell's worked fine.
But if need context, then take the English definition, some dozens of words, paste that into Google Translate, and get back some dozens of words in Czech, German, French, Swedish, Russian, Spanish, Ukrainian, maybe even Japanese, Korean, ....
Besides, for the context of the Web page links, they are nearly always just 1-3 words and really simple, so simple that the translation should be effective. E.g., there are billions of people on the Internet often reading English words for links, and those billions have to include kids in Southeast Asia, etc. with cheap Android smart phones -- they seem to be able to work with English words for links. Right, they also work with icons, but there they get to have their mouse hover over an icon and get the corresponding 1-3 English words.
Net, I contend that for English speakers, or readers, or not, for links on Web pages, 1-3 simple English words are more effective than icons.
For me, with icons I have to hover, and I don't like that; with no icons and the few English words instead, I don't have the delay of hovering.
> You keep saying that and you keep overlooking that a good icon will illustrate the action in a meaningful way.
On "meaningful" I essentially don't agree: For more, to me the idea, claim, suggestion, assumption that icons, e.g., as in the OP, have obvious meaning is outrageous -- I'm torqued. E.g., back to the OP, several icons were shown, and, flatly, bluntly, frankly, I have no, none, nichts, nil, nada, zip, zilch, zero understanding, knowledge, or even a guess what any of the them was supposed to mean -- to me, "meaningful" was just not there, quite the opposite, to me they were flatly meaningless. Then the OP mentioned many dozens more icons, from lots of long, hard work developing icons. Many dozens -- gads, and some people will be pushing those meaningless little cartoons onto me and pressing me to understand them, dozens of them. So, I'm torqued. Like flies at a picnic where I need a spray can of bug juice.
If you see those little cartoons as having meaning, go ahead -- I won't agree. E.g., what the !@#$%^&()_ does a stack of three horizontal bars mean????? What about a tiny circle with a horizontal line to the right over a tiny circle with a horizontal line to the left??? "Meaning"? Not there for me, and I can't look it up in a dictionary, pronounce it, spell it, type it into text (e.g., here), ....
E.g., the Web site of my bank has me log in, and for that they have essentially an HTML "single line text box" control* for my password. Well, bless the hearts of their hard working JavaScript/CSS programmers, inside that text box on the right end they have some icon, some squiggle, some tiny cartoon, some I don't know what the heck. I put up with that meaningless nonsense for some months and on another issue gave some feedba...
Coincidentally:
https://news.ycombinator.com/item?id=30357706
https://www.theverge.com/2022/2/14/22932918/opera-browser-em...
>for instance, singer Kesha’s Yat page is the emojis Rainbow Rocket Alien (editor note: Vox’s CMS doesn’t allow rendering of emojis)
It seems to me as "pure folly", but what do I know, definitely not a country for old men.
For comparison, we pay Hoefler a similar recurring fee for using one font on our website and in our app. I think it's fair. Designing good fonts or icon sets is hard work, and a few hundred euros a year for what's a fundamental part of our marketing and image is a pretty good deal.
UI Icons Line - Pro has more icons than the free counterpart. You can only download low-res PNGs for free, actually. See the part about free license limitations here: https://intercom.help/streamlinehq/en/articles/5354376-strea....
Full details here: https://intercom.help/streamlinehq/en/articles/5354366-strea...
We add new icons and sets every week.
~~That may not be much for the quality you are getting, but my fontawesome infinite access pack was a lot cheaper.~~
Edit: This was wrong. The catch is that you can only download 100 icons at a time, and you can’t download any more when your subscription expires. The system tries to detects bots, so you cannot automatically download the full set.
Our design process is incredibly intensive. https://medium.com/streamline-icons/create-more-beautiful-an...
Which is why a lot of folks choose us over other icon libraries: https://streamlinehq.com/wall-of-love
If you need Purchasing Power Parity, please reach out to hello@streamline.vip.
Right now it’s at the edge of “I need an actual use case for these things, instead of buying them just in case.”
I completely missed the monthly option! Buying them once for $39 and them using them forever feels shifty though. At least if I paid $240 it’d feel sort of fair ;)
What is up with that 100 icons per project though? If I’m paying for icons I want to have unlimited use out of them (especially on websites), not an arbitrary limit. I doubt I’d ever use more than 100 per site, or that you’d enforce that limit, but even if it’s irrational it still bothers me.
I also dislike this is promoted as an open source icon set. This is a commercial product providing free samples. By design, it has a funnel of free / 'open source' users of which some will be turned into customers.
Finally, it's amazing to me as a programmer how commercial minded designers are in tech. It would never occur to me to create a programming library where you have the 'basic' open source version and the 'optimized' commercial version. But basically this is how graphics / icons / page templates / etc are being marketed.
I guess this is where any kind of 'community' helps. If a library has a community, it would be forked the 2nd day if it tries to pull something like this. But icons / templates / graphics are generally the work of a single person or small group of persons. There is no community that may philosophically decide they disagree and thus fork the project on new license terms.
All in all, quite depressing. But, I understand, we all need to eat.
Unfortunately this is frequently the case in programs and libraries: GraalVM has commercial-only optimizations, Qt has commercial licensing and proprietary modules, SQL/NoSQL/caching databases, search indexers, etc.
It's occurred to plenty of others - look at the GPT-3 libraries or a few of the more advanced Deep Learning ones.
Actually, even though training models is very expensive, most models are even available online for free! A large collection is on HuggingFace (including GPT-2, which is essentially a smaller GPT-3), and there are studies proving that the quality is essentially the same. You can literally just download and run them, pretty much like... a free library.
Now I think about it, maybe we should have a blocklist that automatically removes results that invole this kind of deceptive behavior from search results.
We can say that this exists for the wireless networking drivers.
Generally IC is marketed with the features enabled with commercial driver, but shipped with open-source driver.
That doesn't detract from the generosity of giving something away - the only reason it's not been forked and improved by others (for free) is that nobody blinkin' wants to without payment.
Just take what's free, or pay for what's not! I don't know why that's depressing at all.
While the criticism of this being called "open source" is somewhat valid. The suggestion that people creating usable graphics for the community should do so for free is a little frustrating. It's a bit like the memes where designers/photographers/ect are asked to do work for "exposure". The devaluation of creative output is quite sad, people who go out of their way to create something you can use for small fee should be held up as examples of the best of the community.
The vast majority of people who "create" (rather than contribute to) an open source projects are doing so for some sort of "commercial" reason. Wether it's an ambition to built a business around it, promote themselves as an expert, or being done as part of study to further themselves, its all about personal growth which translates to an income.
I fear we are all becoming too used to building upon the "free" foundations of open source and it blinds us to the creativity that goes into these types of projects.
> If a library has a community, it would be forked the 2nd day if it tries to pull something like this
I also find this frustrating as an attitude towards open source, the "open core" concept is completely valid and isn't an argument to fork. With most open source projects the majority of the work is contributed by a very small group of people, true if you frustrate a core contributor they may justifiably fork the project. However we should champion open source projects where the core team find a way to commercialise it in a sustainable way, when done well it only benefits the community.
When picking a foundation to build a business on, the advantages of picking one where there is a commercial organisation supporting the project as a core of their own business model is well known. Examples I can think of:
- Ionic Framework and Capacitor from Ionic
- Tailwind (actually a brilliant counter example with Tailwind UI being the commercial upgrade)
- Wagtail from Torchbox
- Wordpress and Automattic
- Docker
- HashiCorp
It's a weird balancing act. At the end of the day though, everything on the internet is a race to the bottom; nothing beats free. Of course, you don't have to make your work freely available, if you think that your contributions stand on their own merits then you can of course charge a fee. Here's the thing, though: 99% of the time, developers will completely skip over your product even if the free alternative is decidedly worse. For example:
> The vast majority of people who "create" (rather than contribute to) an open source projects are doing so for some sort of "commercial" reason.
Au contraire, just look at Linux, an OS that was designed as a passion project by One Guy and the Internet. It was developed by people who cared, and thought they could create a better system for free. Commercial purposes appropriated it, not the other way around. This is the case for a number of OSS: before people even contribute to an open-source project, it has to be someone's proof-of-concept, someone's toy project. History simply doesn't align with this claim, it takes a lot of Freudian contrivances to make it true.
> I also find this frustrating as an attitude towards open source, the "open core" concept is completely valid and isn't an argument to fork.
Then use your own license. Of course, then it likely wouldn't be considered open-source, but if you're not comfortable letting your community take control of your project then you probably shouldn't use a license that explicitly allows for exactly that. Source-available licenses will assuage your security-minded customers while deterring those pesky contributors and passionate freeloaders from appropriating your hard work.
> When picking a foundation to build a business on, the advantages of picking one where there is a commercial organisation supporting the project as a core of their own business model is well known.
Of course it is. Money follows money. Your priorities as a software salesman are not the same as your priorities as a software user. That disparity is what drives the misery that makes up the modern commercial software landscape. It's the reason why the whole dream of "open source projects where the core team find a way to commercialise it in a sustainable way" doesn't really exist, and certainly isn't championed. Your goal can either be to empower your users at all costs, or to monetize your product. You cannot have your cake and eat it too.
What makes me mad is watching people appropriate the ideals of Free Software for commercial purposes. The only people you're fooling is users, the developers worth their salt won't even give you the time of day if you're peddling some contrived licensed crapware. Every piece of software you listed at the bottom has openly-licensed alternatives, explicitly because there are droves of people who have nothing to gain from paying for the same thing they could get for free.
It's a free world, and ultimately your choice how you choose to license your work. Truly great work transcends these petty concerns with what your users will think and what commercial alternatives exist. Nothing, and I mean *nothing* beats free.
I built wickedblocks.dev ( acquired) and it was completely free.
Some stuff, were better coded than TUI.
Definitely not an upgrade,...
That comes from the roots of programming and computer science in general, where it's part of the old-school "academic spirit" is making stuff available for others to experiment with and advance on. Universities paid their professors, doctors and other academics to work on all that stuff like compilers, programming languages, communication standards or operating systems - wherever you look in open source projects dating prior to the 2000s, all of them have a strong foundation somewhere in academics or the military.
Only when computers became household items, commercialization really took off, as suddenly the market for software was no longer restricted to fellow academics, a few very ambitious hobbyists and even fewer megacorporations.
UI/UX design, in contrast, never had that "golden age" of third-party funding everything that the people wanted to work on in that moment.
Source code is here https://github.com/webalys-hq/streamline-vectors/blob/main/R... and in https://www.figma.com/@05466272_9382_4.
Thank you for the feedback.
Publishing open source is like me publishing only the binaries of my tool and thn saying because it's free it's also 'open source'.
So, assuming you used some tool (Photoshop, Figma, etc) to export to SVG, open source would mean you publish the original files too.
> The "source code" for a work means the preferred form of the work for making modifications to it.
Unless the designers from Streamline are working directly with the SVGs, which is possible, but unlikely. To be truly open-source, it would have to include, for example, Illustrator files.
Otherwise, it is free, libre, permissive or just open. But I wouldn't call it open source unless we have the source.
Open-core (open source with enhanced commercial version or commercial first-party add-ons) model is not uncommon in software (libraries, applications, etc.)
Neither is having a downstream commercial derivative vendor being a major contributor of work and resources to an open source project, which is essentially the same thing. (e.g., EnterpriseDB & Postgres.)
It's not something that the design industry does but software doesn't.
If anything open source should encourage early open sourcing, doing work in the public, accepting contributions, etc.
I think it's more like a shareware, ie. not open source.