Is Semantic UI development dead ?
GuacheSuede opened this issue · comments
@jlukic 's last commit was 5/6 months ago and even then did not have many improvements nor much changes.
What would be the status of Semantic UI especially with Bootstrap having much more development than this ? Is it in active development or are any features planed for the near future ?
Note: This open source project goes dead without @jlukic or @quirkyjack , zero developement, zero updates, and for nearly a year at that. Even pull requests are not moving. Could there perhaps be a solution to this ?
Please do not say plans for something, for that might be for another half a year of waiting with no promises or answers
There will be a meeting at 6PM EST tomorrow on Hangouts !
I hope to volunteer to maintain/run this project if @jlukic does not have the time for it
Many many people are same interrogation : Semantic UI is great framework (with default but lot of good points). I'm really sad to see development stopped. For development (newer), using a framework without news about futur is not attractive.
Same here I'd love to see semantic continue its development.
Maybe the semantic team works on something huge, like a V3 ?
Nothing since august last year
I've decided to move to material ui, this isnt worth investing time
I would however be happy to pay an annual fee for a premium plan that will guarantee bug fixes and improvements.
@jeveloper I agree. perhaps a fork could be created
Agree too. I'm developper and do little stuff for fun. This month, I begin new project (not for work) and have to choose between SUI and Metro-UI CSS V4 (I'm already using V3). Finally I choose Metro-UI : lot of high level components and very active developper.
Too bad : I like SUI, it's very good framework but no dev and with time, chatroom become empty... so no support.
For fork, possible if people are ready to participate...
In my opinion, SUI is almost perfect ^^
But :
- be more easily customisable (skin) : actually it's difficult to modify color or add background on component (or I'm bad... possible because newer with SUI ;-) )
- several components missing (need 3rd party now) : date / time picker or have image slideshow.
But it's my feeling after my poor SUI experience, why don't making "BOX IDEA" here or elsewhere ?
PS : sorry for my poor english.
I have been considering making a fork for a while now and I would love to help get peoples features added and patch the current bugs.
I have some knowledge for the semantic code base but I will have to learn some more before I make a fork.
I will 100% make a fork if people are willing to contribute. 😃
@hammy2899 Count me in, I think it's doable.
@GuacheSuede @hammy2899 @arjarn @jeveloper @etshy Please read the following:
- https://gitter.im/Semantic-Org/Semantic-UI?at=5a3ca96fffa3e37919693fde
- https://gitter.im/Semantic-Org/Semantic-UI?at=5a3caa595355812e5705bfe2
- https://gitter.im/Semantic-Org/Semantic-UI?at=5a3cabadc072deaf0bee5281
- https://gitter.im/Semantic-Org/Semantic-UI?at=5a3cad03ffa3e379196950bc
- https://gitter.im/Semantic-Org/Semantic-UI?at=5a501e8c03838b2f2a7760ae
@Bodom78 Could you share what great concepts Semantic UI has from your POV ? For me it is mostly Modular Theme-ing
As there seem to be people willing to support SUI, maybe @jlukic considers adding them to the team.
@GuacheSuede The human friendly markup is something I love, I find much easier to work with compared to Bootstrap or Foundation.
Another thing I really like is the amount of variations on offer for elements, look at the lists or buttons pages in the docs. So many variations already there and with the theming, overrides and variables it's easy to add your own building up on what is already there.
I've found I can take a website from design to html much faster and with less nonsensical markup then the other frameworks I have used.
As a start I would love to see more contributors working on clearing out the issues and pull requests.
@witrin : thx for information
+1 @Bodom78 : like SUI for same things, can add bricks to make great site with original design, but I haven't find easy way for theming so for this, I'm agree @GuacheSuede
And, obviously, priority to resolve issues.
In last, more high level components : I think, dev / people like have all-in-one framework. Tiresome have 4-5 third party library for image slide, date time picker or other...
I know, 100% needing is not possible but several components missing in SUI in my mind.
That is good to hear ! @Bodom78 @arjarn
@Bodom78 I agree, SUI's feature of markup seems much easier than Boostrap or Foundation. I agree that the variations of the the framework seems wonderful !
@arjarn Thanks for agreeing :). I agree SUI is lacking a few features that would be much desired. Could you list out the components you would hope to see in Semantic UI ?
Issues are a large priority.
I am thinking it would be better for Semantic UI if it was switched to SASS/SCSS due to more features
I have so far not been able to get @jlukic reponse when I private messaged him a few days ago
Since the discussion includes forking, I suggest that future maintainers seriously consider moving Semantic-UI to CSS Grid.
@Joshfindit Definitely would, do you have an example of a framework that does this ?
Here are my two cents:
For me it's the core concept of this Framework what makes it so valuable and (of course) the rich feature set.
We are using this framework with our custom build tool chain (YARN only with Fluid templates) and the big effort with this concept is separation of concern! We don't need to copy fat boiler plates for each project, we reuse our investments, we spent less time and the projects are just that big they only have to be. So the possibility to layering components with definitions, themes and sites are a huge step forward in this (CSS) area.
Also having a clean wording makes it more fun to work with this stuff ... but I know structures needs to grow and kept clean, so the only thing I'm most aware of when a fork would came up, core concepts and/or structure became a mess because too many cooks (might) spoil the broth. E.g. when tickets like might accepted in a hurry #5785.
I also think the @jlukic should be still part of this; sure, if he want. Maybe he could find a way the community could be involved a little bit more, thus we get faster releases and he gets more free time. I also see financial support for this project should be a part of the future discussions.
@witrin Yes definitely.
I agree too many cooks (might) spoil the broth and that @jlukic should be part of this as well as financial support as an issue to be brought up.
What I am suggesting however would be a change of operation for the project has come to a halt for a year and it would be a waste if it died especially as you said of the good concepts of it (modularity, build tool chain, rich feature set and wording)
I assure you that I totally agree with you especially on the key point of keeping structure and maintaining quality.
But the dilemma is of such that Semantic UI is not moving with @jlukic nearly disappearing for a year, remember this is a 39 thousand stars project.
Do you think a Semantic UI should stay dead ?
@GuacheSuede Semantic UI should definitely kept alive! :D I think the next step could be a discussion about a change of operation for the project; like you mentioned. If I get the last messages from @jlukic right, I think he is currently in Russia on vacation. So maybe we wait a little bit more until he pings back and a discussion about operation then could be forced. And when he really would not answer within the next n weeks (what I don't think) maybe a discussion about a community driven fork needs to be started earlier.
@witrin Yes, I agree. Glad we are on the same page and understanding, perhaps let's wait till Mid Feb and see if he responds till then ?
Note, Semantic UI React is still in development. I'm the owner/author over there and I have owner rights to the Semantic Org at large. I've also had offline chats with Jack about the future of SUI.
Jack's focus is mostly on pursuing a semantic component standard for the web at large. A W3C for UI, if you will. Frameworks like Semantic UI and library implementations like Semantic UI React would serve as premier examples of how to implement the UI specification.
All said, I am interested in keeping development moving forward on both Semantic UI and Semantic UI React. The next frontier, in my mind, is to abstract out the styles and state of the components so that other implementations can more easily adapt them. It would also serve to protect us against future changes in the technology landscape. I've discussed this with Jack on a video call and he is in agreement, regarding the abstracted styles at least.
In my experience here at Semantic Org, what we need is a core of truly dedicated members. Over the last few years, I've seen a lot of conversations like this one begin, and sometimes an effort is started, but there is no follow through. I think the reasons behind the lack of follow-through are what we need to address.
One idea I've had is to start with regular meetings. At least get a core group of folks together to talk about a realistic way to move forward. I am more than happy to hear thoughts as to how we can get folks to commit in a sustainable and productive manner.
Who's up for it?
I have been in contact with @jlukic for the past couple of weeks and hopefully there will be a patch before the 1st of Februari. He has been quite busy lately.
I would be more than happy to help in any way I can!
I think it'd be reasonable to have a video call with people who are interested in SUI to discuss what's going on in the project.
I'll summarize the story of the project here, as I know I've told it through different mediums at different times and it's probably more difficult to track these days.
My goal is to share my thoughts as I understand them myself, naked without a filter, and I hope you can do me the kindness of being forgiving for any parts where I may have faltered in character or conclusion.
I've generally thought of the core SUI goals as a descriptivist standard for web UI, a counterpoint to the top-down approach of the W3C. This is something that is still my primary life ambition, and that I hope I can be thinking about until my later days.
The issue though, is that indefinitely financing open source can be difficult. It, at the very least, requires enough financing to cover one persons expenses indefinitely which, in real financial terms, is a multi-million dollar affair. I'll get to that in a bit though. Perhaps first I'll try to share the story of where the project came and where I am now.
Semantic UI 1.0 was lucky enough to exist because I was able to open source much of my work done at the now defunct startup Quirky where I was working as head of front end development on solving the general issue of how large teams can work on a front end codebase without developer scaling issues. The aspirations for the project were to show how aspects of semantics from natural language, (noun modifier relationships, plurality, significant word order) were better at describing physicality than classic CS models which function more like 'cooking recipes' (or behavioral specifications).
When I left Quirky I ended up travelling a bit and working on Semantic UI, generally spending through savings and hammering out ideas around types of UI, naming conventions, reasonable methodologies for constructing UI definitions and generally looking for a path forward.
Around a year into this plan I got an interesting email from someone who had recently bought The New Republic. I ended up coming in and discussing what I was doing, and found out it was a good match. I was offered an interesting arrangement where I could work to help recode the magazines aging codebase (using SUI) and redesign the magazines website, on a 75/25 basis while working on SUI one week a month without pay out of their office.
After around 6 months the redesign was done, and I was itching to get back to SUI full-time. Chris was kind enough to offer me the opportunity to keep a desk at the New Republic offices while working on open source full-time. I then spent the next year, keeping a desk at their Union Square office, greeting guests as if I was an employee, but working without pay on open source full-time--from summer 2014 through summer 2015, when i finally finished 2.0 of Semantic UI.
This was my second period working entirely from savings and taking no salary while working on SUI. In all honesty, it was perhaps the most fulfilling and liberating period of my life. Though at the end of it, on medicaid with no money, I knew it wasn't going to be sustainable. I had now saved and then spent ~$60-80,000 in savings in two separate periods working on Semantic UI.
At this time I started talking to some investors around perhaps raising a seed round for a company supporting Semantic UI. From talking to investors and angels, I was introduced to some interesting people who founded other OS companies like Meteor, and Ionic and had calls to discuss their story. I also started talking to a list of valley VCs, preparing my seed round slide deck, and began to get ready to sell the idea of Semantic UI as a company.
But something hit me in my gut, that felt wrong about it. When I think about the people that have inspired me to create in my life, the idea of attaching a 'profit motive' created cognitive dissonance. There was no "Dostoevsky LLC", or "James Joyce C Corp". Perhaps, one could argue, there were publishers, but these arrangement didn't require the same quick turnarounds and preposterous hockey puck revenue charts as is expected in the valley these days.
The idea of attaching a company, and shareholders to SUI didn't feel right to me, and from looking through the direction of open source companies who had tried, it did seem clear all of them, without fail, ended up following the money eventually. At its worst, if I relented and tried anyway and decided it wasn't correct the intellectual property for SUI would, most likely, never be available again--owned by the stakeholders in the company.
While I was weighing these thoughts, a few SUI companies were reaching out about 2.0 and how to use SUI in their company. One of those companies Qualia was starting to raise their series A and asked me to come out and look at their codebase and give feedback on their design. They were fresh Stanford and Penn grads using an entirely Meteor/SUI stack, and with very stimulating views on code and business. From talking with them, it was clear I would be able to 1) have access to a significant equity stake in an early stage company that had not been available in any other period of my life 2) be able to use SUI as the core of a startups codebase and test out longitudinally it's utility and triage any scaling issues.
Around fall 2015 I ended up signing contracts and starting full time as head of design at Qualia, something that I was very sheepish to talk about publicly because it was a difficult conceit to explain. But I was penned in for two years, which seemed in the scope of a lifelong project as "not forever", and so I soldiered on.
I ended up agreeing to a "week on week off" schedule from NYC -> SF. Which in hindsight, may have been a very silly thing, leaving very little weekend time for open source that wasn't recuperating or traveling, or reconnecting with friends. My 2 years ran up in July of this year, and I was offered the same equity to stick another 2 years, a contract I accepted.
Now it's not difficult to see how this could be construed as 'selling out', but having gone through my life savings twice, having tried the contracting/OS balance, having looked at SUI as an excel spreadsheet of a seed deck's financial projections, and having calculated the multiples of current donations that i'd need to receive to be able to pay rent, it didnt seem feasible to go any other way.
The only way forward I can imagine for the project is to raise, at once, enough money to carry the product indefinitely, what I've been thinking of as 'escape velocity' for OS. This number begins at the cost of one developer to work, in perpetuity, on an OS project. The calculations for this looks something like this, the numbers may vary somewhat, but coarsely still look the same.
Cost of living NYC
$85,000
Cost of living with savings and possibility of family
$125,000
Safe rate of return for annuity/investment
4%
Endowment/Annuity that could produce cost of living indefinitely
$125,000 / 0.04% = $3,125,000
Long term capital gains tax
20%
Amount of pre-tax money necessary to finance an endowment for one-man Semantic UI non profit:
$3.906,000
All and all this is a fair amount of money that doesn't seem reasonable to raise without institutional investment, or personally putting up enough capital to cover the lion's share of the non-profit.
I'm trying my best to do the latter for now, which has caused me to make some clear and perhaps single-minded concessions for the project.
Generally, I think this idea of developer annuities, is something that I will be trying to pitch as a means of escaping the treadmill for OS developers. I think the numbers can vary between $1-4m per-person in funding, but paint a more clear picture of how much it truly costs to support something which can exist in perpetuity. When not thought of as personal wealth, but as an endowment, or non-profit, perhaps it's not seen as unreasonable. I've been thinking of this as the "wellspring" model of open source, and if it's something I can prove for myself, I'm hoping I can use it to help other projects finance their visions with the long-term in mind.
All and all though, if I'm capable of financing a non-profit, where I can say for certain my children will be able to have long term access to an open source UI framework (or standard) without profit motive, or need for solicitation, I'll be very proud of this time.
@GuacheSuede: Currently there is no framework that I'm aware of that is based on CSS Grid. It's very new (but has excellent browser support), and I think this is a technical opportunity as well as an SEO opportunity: CSS Grid does a lot of things very intelligently and so obsoletes some of the old methods of page placement, and for as many people who talk about 'with CSS Grid we don't need [Bootstrap] any more', there are possibly more who recognize that CSS Grid doesn't bring design cohesion, design language, or even a design direction and these are really where Semantic-UI shines.
@jlukic Thanks for the update, it's good to know and have put into perspective the origins, difficulties and future wishes for the project.
@levithomason If the project was to move towards adding some dedicated contributers to assist in maintaining the project, then I feel a "call to arms" would best be placed on the semantic website somewhere. I generally find Git issue threads is mostly users looking for fixes and messaging can get lost easily. (Looking forward to using Semantic UI React in the future!)
If you'd like to participate in a conference call about Semantic UI, I've created this link for listing availability.
@jlukic Thanks Jack for your long and candid explanations in relation to funding and your struggles.
I think that we need to talk yes and personally would like to manage the project if you do not happen to have time for this as i do see many new features that can be done that would benefit users.
I have filled the poll and look forward to talking to you soon.
@jlukic It's nice to hear any updates about SUI because even without actively maintaining SUI is the best OS CSS library available, it's much easier to use than any other CSS library I think it would be nice to have more maintainers to review all pull requests and issues. Maybe you should look to React SUI, they tried to solve the problem of styled components, but without a result.
For what it's worth i jammed out a release tonight
https://github.com/Semantic-Org/Semantic-UI/releases/tag/2.2.14
Nice to see motivated people for this great project.
Agree with @jlukic's thought about SUI.
Unfortunately I haven't lot of time to devote for new project.
But I'm coming here for proposing ideas or issues ;)
Good luck for conference call, I cross my fingers for positive issue :)
@jlukic I would be happy to help out as well. Unfortunately I wouldn't be able to join on those meeting times. Simply because I live in Eastern Europe and for me, those times are from 1am to 5am.
Hey folks!
Been using Semantic-UI as my projects' framework for a long time.
It's good to hear that is not dead.
Thanks @jlukic for your honesty and of course for sharing your time and work with all of us.
Hope that you guys can manage to keep things going.
@jlukic It's great to hear what you're trying to do with the project and I fully support that and would love to see semantic grow into a standard for modern web.
Over the past few weeks I have been contributing and helping out by updating some things and I want to help out as must as I can because I love this project and I have been using the framework for well over a year now, with that said I am hoping to become a contributor/maintainer so I can help sort out the issue right now with the current amount of issues/pr there is and get semantic in a stable bug free version with components for every scenario while keeping the style and easability which Jack is going for.
I have joined 2 of the days on the meeting planner which @jlukic posted so I can see what the next moves are and so I can help out in anyway I can.
While I cannot make the discussion, I'll offer my thoughts here beforehand.
A community seems to be here to help. How do we capitalize on that desire for the health of the project?
- Open communication - this is critical for enlisting the help of others and "marketing" SUI for both funding and community help
- Small, focused tasks - these can be done for "free" when distributed to many contributors (this is more possible with tests; e.g. visual regression tests)
- Management and direction - these become a primary task of either the financially supported leader and/or other leaders of the project
Things I would like to see in the future:
- Making SUI Javascript framework/library agnostic - I recognize work is already going on to do this
- Incorporate features (e.g. calendar module) proactively - the community has already written them
- Regular releases - technology changes (e.g. CSS grid, drop IE support, etc.), bug fixes, merged features
I think these things are possible, albeit, slowly, regardless of funding if there is by-in by the current leaders of SUI.
@khornberg : nice analysis.
I cannot make the discussion too (schedule problem), so also offer my thoughts here...
Calendar and slider module are already written (for slider, I think it can be improving).
Be great to incorporate it in SUI.
For slider, technically, why not same as Vegas Slide (https://vegas.jaysalvat.com/) : using background image in HTML component with optionals controls to swap between pictures. With that, it's possible to place above picture text, button... offering lot of possibilities as mostly SUI components ;)
Maybe others components can be adding :
- Toasts to display informative message (less perturbing user than modal message), like this http://materializecss.com/dialogs.html or https://metroui.org.ua/v4/toast.html
- Chips like http://materializecss.com/chips.html (like dropdown SUI)
and for design, wave effect when you activate button (like https://mdbootstrap.com/javascript/waves-effect)
All that is I'm dreaming in my devs : I know some ideas may be good, other very bad ^^
Last point, with multiples / heterogeneous device today, it may be useful to add two more breakpoints for responsive : one below 768px (576 ?) and one upper 1200px (1500 ?).
@arjarn I agree with getting the calendar and the slider (range) module merged into the core UI and I also agree with the idea to add toasts or some sort of notification yet I don't agree with the button ripple effect, this is a material design feature and in my option doesn't fit the design pattern for semantic. I think adding lots of features like this would make the framework too clunky when it is only required for that specific thing therefore you could just add it to your theme.
I definitely think a talk needs to be had about new and current components on what needs to be added and what needs to be changed to stay relevant to the current web standards and design.
I also agree with adding a new break point for smaller devices like smart watches for example.
@hammy2899 I agree your opinion about button ripple. Perhaps not so good to mix components features and design feature.
@arjarn I can see toasts being handy, right now we made a custom saving notification that served that purpose but having it built into the framework would be really nice
@GammaGames Definitely sounds good !
What about a KickStarter or similar where some of the more desired components would be worked on as part of the reward? Font Awesome raised millions in the last year for version 5 of their product... maybe something like that could work for Semantic?
We just released the facelift of our new automotive classifieds website based completely on SUI. We definitely want to see this project continue moving forward. Once the site gets a bit more traction, we'd love to help out financially.
It seems 3, 6 and 7 Februari are the dates most people are available for a conference call about Semantic UI.
The most people can do the call on the 3rd, I will be out of country that day but I can make some time available to participate.
@jlukic are you organizing the call?
Perhaps an endowment wouldn't be necessary. ~40,000 stars x $5 / year. I'd sure be willing to do that for a LONG time.
What is the opinion of core devs on a kickstarter? font-awesome kickstarter seems to have done well for them.
If there was a kickstarter it would have to be done well, but another alternative could be patreon. Just throwing some ideas out there for the people making decisions on the project
+1 for patreon,
@yyx990803 (Evan You) main developer of Vuejs framework, use Patreon to dynamise developpments of VueJs
Perhaps a way to provide ongoing support would be to have an online SUI theme builder that would be paid (seems like most online theme builders are paid - or else the theme builder is free but you pay for the library). I realize that this would require additional coding to put in place, but it could also be a tremendous help in spreading the use of Semantic UI. A Kickstarter/IndieGoGo would be best for a project like that, of course. A Patreon campaign could be started in the meantime if necessary. I'm in for any of these ideas including simply long-term support through Patreon.
@DanAtShenTech I concur, will look into it. Depends if @jlukic has time, else i hope to maintain it
I've closed out the poll for the conference call.
The date is set for 6PM EST tomorrow
I'll post the hangouts link in this thread around that time, so anyone who is around can join as well.
Semantic-UI is probably the best CSS solution for now. It is much more flexible and elegant than Bootstrap which I had used for a long time before meeting this masterpiece.
Recently React is becoming a game changer for this framework. No wonder since many of us know how good is it when used as a React UI library.
Semantic UI for React trends
NPM downloads
Considering popularity growth it is crucial to continue the development. Hope that this conf. call will help to find a reasonable solution.
Do we have an agenda for this meet?
I hope SUI can keep going. It's the best I've found (especially for Rails with all it's dependencies)
and my default framework for all new projects.
This week I decided to play around with Rails 5.2 for a new codebase I'm tinkering with. SUI is the only one that came up clean the first time without needing any major tweaks or configuration changes. Semantic wouldn't fire up at all - none of the JS worked. And Bootstrap required a bunch of other jQuery gems and stuff to get working - plus it's on the fat side, and doesn't look as clean as SUI.
SUI really is a hidden gem. A shame more people don't know about it.
Please record the Hangout session on youtube. Sure many of us want to know what was said during the session.
@menthol that might not be a bad idea.
Gitlab live stream a lot of their meetings/discussions. Gitlab's YouTube
I'm not sure I have an agenda, does anyone have suggestions. I'll look into recording as well
I know we'll discuss the long-term and grand vision for SUI. I would like to address some of the shorter term goals and current issues. Being the React integration author/maintainer, there are things I'd like to do there in terms of themes and styling that I can't reconcile entirely with SUI.
Jack, this is similar to what we've talked about before with abstracting styles into a central location so that we can generate styles for all integrations. There are some other implications I'd like to consider here as well, such as removing child based selectors and isolating styles between components (e.g. a label in a menu item wouldn't change its style).
I think my focus is primarily on where we are now and how we take the next immediate step. Unfortunately, I've become quite ill and I've completely lost my voice 😕 There's a good chance I won't be able to join.
If you, and the others, could address a feasible path to moving forward with what and who we have right now, that would address what I'm most focused on. The other option is to fork Semantic UI React into a new effort that handles these things, but I think it would be a shame to split what little forces we have already.
@levithomason you can just listen in i would hope, your insights would be valueble
I will do my best 🙇
What is the link for the hangout?
Just figured out youtube recording
https://hangouts.google.com/hangouts/_/kxc753flhfdfxmz75ifrxymrbye
@DanAtShenTech @levithomason @GuacheSuede @hammy2899
I'll hang out here until maybe 6:15 before getting started
How was the meeting? Any decision regarding the future of Semantic-UI?
I believe the results were recorded on YouTube. It was at the least broadcast live. The results were promising. Jack took some good notes and recorded some actionable items. The high-level picture is that steps will be taken to reduce the number of open issues and keep progress going while the larger funding issue is tackled.
@levithomason Do you have a link ?
I don't, I just saw the live indicator in the upper right of the call.
@jlukic Can you provide the link of recorded hangout?
No feedback about meeting ?
Short summary will be great ;-)
Any news ?
From Where I stand, i have reason to believe Semantic UI is DEAD.
Jack has not replied my many emails or gitter messages despite promising to in the Hangouts chat
I feel that nothing was highlighted nor any plan was created during the Hangouts chat ( I was there ) and
Jack has not come out with any funding plan. in fact he refused due to his desire for his "idiosyncrasies " to stay in the Framework
He has also highlighted he himself does not have the time to maintain this project
Would anyone be interested in a community fork ?
A YEAR of no substantial updates, those who disagree don't think it's dead ?
As it as already been told, there seem to be people willing to support SUI, maybe @jlukic or @levithomason considers adding them to the team...?
From what I gathered Jack made some notes from the hangout and did start making some sort of a plan and he did agree to add the people who where in the hangout to the semantic team to help maintain it.
I think we need to wait and see what Jack says before we make a fork because we don't exactly know why he isn't responding, anything could have happened.
Lets wait another week or so before we start planning a community fork.
I would be up for helping to maintain the project. I and probably many others would love to help but at the moment it feels like not a lot of progress is being made and there is no good way to help the Semantic UI community.
In general Semantic-UI seems to both offer unique features and try to do other things that other frameworks such as Bootstrap, Bulma, Foundation, Polymer, etc. already specialize in doing better. I would be much happier to see Semantic-UI focus the precious resources of its community on the unique value that it can provide.
@hammy2899 @brodybits @dbendaou
- 2.5 weeks of no activity since the chat (we already opted to wait a week... 3 weeks ago.)
- Month of no "real" updates since I started this issue
- YEAR of no updates in the project as a whole
For clarity, SUI development is not dead. It is, however, quite slow. You have to realize what it is like to have a repo with 40k stars. Jack had somewhere around 8k Github notifications on the call we had. It is unrealistic to think he can address every request personally. Please realize, we both have full-time jobs that take first priority of our development time.
It is also safe to say that SUI is not currently undergoing drastic changes or improvements. The React port is approaching a similar position, though, continues to iterate. In the background, I can say I have some experiments going that may influence the future of SUI's theming system.
I would forgo public opinion in favor of data regarding progress, though. Here's the last month's worth of data on development in Semantic Org. I've also shown user adoption continues to increase for both projects.
Repo Pulse
Semantic UI
https://github.com/Semantic-Org/Semantic-UI/pulse
Semantic UI React
FWIW, the React port which I manage has the lion's share of development and progress these days:
https://github.com/Semantic-Org/Semantic-UI-React/pulse
NPM Downloads
Users aren't slowing down on package installs either.
Semantic UI
https://npm-stat.com/charts.html?package=semantic-ui&from=2015-01-01&to=2018-01-31
Semantic UI React
https://npm-stat.com/charts.html?package=semantic-ui-react&from=2016-09-01&to=2018-01-31
Not sure why this tool reports 173k, we're actually pushing 200k downloads a month.
It is easy to think of software like other paid products and services. However, OSS is free and it is often borne out of the passion and sacrifice of a small few and offered to the world for the sake of doing exactly that. Please try to remember that when setting expectations and making requests.
I will leave with a call to action. I started developing on SUI in 2015 out of my own needs. What has astounded me is the proportion of users asking for more and not giving back. As a maintainer, I'm always hoping someone will jump in and join forces in the mission, making a better web framework. I'm beyond grateful for the few that have, but I dream of a day when initiative, contribution, and perseverance outpace the barrage of requests. Short of personal responsibility and action, we'll simply have to wait for Jack to figure how to arrange for indefinite payment for himself and others to bring improvements full time. So, roll up your sleeves, fork repos, submit PRs, assemble issue triage teams, engineer solutions, and build the thing you are waiting for. The world is before you and it is yours to take. We're here.
Reading back, that sounds a little harsh and whiny 😸 Totally not my intent, just trying to shed as much reality as I can on the topic.
@levithomason I entirely agree with your call to action. As a maintainer, as one who can merge things, none of "fork repos, submit PRs, assemble issue triage teams, engineer solutions, and build the thing" can happen unless people (the maintainers) can and do merge.
However, merging is simple. Making sure the software works as intended afterwards is difficult. So communication between PRs, issues, ideas, requests is necessary. Someone must coordinate all that. That coordination cannot happen unless there is knowledge of the plan ahead, unless there is a purpose, a goal, a destination to reach and plan of how to get there.
From the start of this thread, my desire is that those questions are answered. Without it, SUI may disintegrate, albeit slowly.
So:
What is the value or unique quality (as @brodybits mentioned) that SUI provides? What distinguishes it from the other UI frameworks out there? What is SUI's purpose?
Then:
What is the plan henceforth? How do we accomplish SUI's purpose? How do maintain a UI framework in a world that changes?
As a community or as maintainers or as the creator, when those questions are answered, a plan can follow. Once the plan is communicated, the few and precious resources that others can offer can be focused and used wisely.
Agreed. I'll do my best to answer from my perspective. This also prompts me to discuss what I think is an upcoming split between SUI and SUIR.
What is the value or unique quality (as
@brodybitsmentioned) that SUI provides? What distinguishes it from the other UI frameworks out there?
For me, it is the semantic design language and it's thorough application to components. Explicitly defining component types, content, states, and variations and applying those in matrix fashion across all components makes SUI one of the most feature complete frameworks there is. In the React port, I can often change the name of a component only, leaving all props and values as-is, and the get a brand new experience that still works. This is because all components share this common language and apply it effectively.
Second, theming. Although the current process is a royal PITA, the power of SUI's theming capabilities is unparalleled. This is why I hope to help reign it in someday with a more user-friendly and shareable theming experience.
These core values will follow me everywhere I go.
What is SUI's purpose?
SUI
Jack has answered this better than I can. In an attempt to summarize, Jack's lifelong pursuit is something of a W3C for UI. A standard for the language of the web components. SUI itself is something of a paramount example implementation of what that might look like.
SUIR
I created this strictly for business purposes. I needed to build apps with a highly themeable framework that included as many components and variations as possible to avoid reinventing the wheel. This is also why we continue adding "addon" components to the library that don't exist in SUI core. It's purpose is simply to be the best React component library available. Best to me means tests, docs, maintenance, and a plethora of components and capabilties.
What is the plan henceforth?
SUI
We're pending Jack's response here for SUI, but he indicated he's still looking for ways to pay for development.
SUIR
My aspirations and plans are a little different than Jack's. I plan to continue tackling immediate business needs for component libraries. In fact, I'm starting a new job at a prominent employer and will be tasked with doing exactly that on a global enterprise level very soon. As such, I will either be making drastic updates to SUIR to support this mission or (more than likely) I'll drop SUI's CSS and ship a new component library heavily inspired by SUIR.
How do we accomplish SUI's purpose?
SUI
Again, we're awaiting Jack's notes and direction for SUI core.
SUIR
Our purpose is to design the most friendly and powerful component API there is. We have a very clear README section outlining exactly how you can help. We have specific issue labels identifying issues needing comment/feedback, needing engineering, needing general help, and good first issues for those getting started. We also have an extensive CONTRIBUTING.md detailing how to create components and use our stack.
How do maintain a UI framework in a world that changes?
This is the key question for me! How do we not constantly rewrite so much code everytime technology moves on? How do we share components and styles between static pages, jQuery sites, React apps, native mobile apps, and the next JS framework to overtake the community? How do we create a framework that is simple to theme and can be used by any business or enterprise?
These questions are exactly what has pushed me to my most recent conclusions and prototypes. Thank goodness, they are also the reason for my new job opportunity where I'll get to build them out.
I believe the answer is to abstract and generalize as much of the duplicated work as possible.
-
Shared styles. SUI CSS is powerful but it is also limiting, IMO. It is written in LESS and only works in a browser. Also the pervasive use of child selectors and important make it difficult to integrate elsewhere. Following the SUI design language, create a framework agnostic style framework that can be used anywhere and compile down to any format. JS is the best candidate for this in my mind.
-
Shared state. SUI and SUIR have duplicated control flow for the Dropdown component with thousands of lines of JS each. This is a tragedy. Once Vue is the new hotness, we'll do it a 3rd time. That is insanity. Many modules are well represented as finite state machines. I want to further explore this and how they can be mapped to frameworks.
As a community or as maintainers or as the creator, when those questions are answered, a plan can follow. Once the plan is communicated, the few and precious resources that others can offer can be focused and used wisely.
Exactly. I can't speak for SUI core as that is Jack's alone. Part of my new role may include building this framework as I'm describing, evangelizing it, and building community. I can't guarantee that nor that it will be soon if it does happen, but it is at least the meeting of minds we have at the moment.
Thank you for your detailed answer Levi.
I will leave with a call to action. I started developing on SUI in 2015 out of my own needs. What has astounded me is the proportion of users asking for more and not giving back. As a maintainer, I'm always hoping someone will jump in and join forces in the mission, making a better web framework. I'm beyond grateful for the few that have, but I dream of a day when initiative, contribution, and perseverance outpace the barrage of requests. Short of personal responsibility and action, we'll simply have to wait for Jack to figure how to arrange for indefinite payment for himself and others to bring improvements full time. So, roll up your sleeves, fork repos, submit PRs, assemble issue triage teams, engineer solutions, and build the thing you are waiting for. The world is before you and it is yours to take. We're here.
Back in July I submitted my first pull request to Semantic UI (#5507). At my job we were using my forked version of SUI until it was fixed in 2.2.14, for which me and my colleagues are really grateful.
In the period 5th of July - 19th of December nothing happened with my pull-request. On the 19th I started having email contact with Jack to see if my fix could be merged and if there was any way to help him out. This ended up in Jack promising a resolution by feb 1 which resulted in 2.2.14.
I understand how difficult it is to have a fulltime job and maintaining SUI. But at this moment as a developer I don't have a good reason to be active in the SUI community. I see pull requests being made every week that will either not be merged at all or it will take months before they will be added in a new version.
@dannyBies I agree we do need some more people who can approve PR. Jack discussed this in the hangout of possibly making PR need 2 people to approve them before it can be merged and I think this would work really well but SUI is lacking maintainers with said power.
I also agree. I don't use the majority of SUI, just the styles, as the React port includes all of it's own JS and renders it's own HTML.
That said if I were in this position with SUI, I would consider forking and maintaining that. If I truly thought I could do a better service, I would do it. When someone else came along (possibly Jack) and resumed providing a better service than I was able to, then I'd gladly merge and join forces.
This is actually what happened in the Node.js community with io.js. There were some differences and disagreements, so io.js forked from Node.js and took off. After considerable progress, everything was resolved and io.js merged back with Node.js.
I guess I'm giving another call to action. Don't stop at submitting a PR. Just solve the problem in front of you. If that problem is PRs not getting merged, figure out a way to merge them. Creating a fork is one way. If there is not a clear plan for making enhancements, fork the project, create a plan and invite everyone to join you. The world is truly here for your taking.
Consider all the active people in this thread who feel as you do. It seems to me they just need a leader to pull it together. Why not be that leader?
I know one hurdle to merging PRs is getting thorough and proper reviews. I wonder if we could set up some kind of process or tool to merge PRs based on serveral reviewer approvals.
I do have this problem at SUIR as well. We've added a number of contributors to help but it would be great to leverage more of the community for this process.
I imagine having a core group of folks who can skim and sign off on code quality and then open to the community for reviewing and approving the feature/fix itself.
Forking should be a last resort as splitting an already small community will most likely end in disaster. I do think the only solutions are to fork or to add a group of people chosen by @jlukic to help maintain the project.
I say we wait until the 25th to see if we can find a solution and if not, we fork with the idea that we can merge back to the main repo when we do find a viable solution.
I do not have a lot of time to be the owner of the fork, so if someone else has the time and wants to do it (with me) comment here. If not I will fork on the 26th.
@dannyBies I agree and would be happy to help you maintain the fork or be the owner if you don't have the time.
I would start with small bug fixes only and ship a new npm package to users can install it. This should make it easier to merge back to the main repo later. Also, staying current with the main repo will be important.
My hope is that a fork would produce a team of maintainers and prove that bug fixes can be shipped reliably. Perhaps this will be enough of a signal to allow those maintainers to merge to master on the main repo.
Bulma is nice, but it's not as complete a solution as SUI. You need to add in form validators and something like Masonry and a bunch of other stuff to match the overall functionality. Bulma is pretty strict about "mobile first" too - building a robust Rails app is kind of a pain when you have to keep trying to work around the assumption it makes.
I'm really sad to read this. In my mind SUI (without "R" ;) ) is a great and unique framework..
Multiples bricks we can use to build various site easily. I never see this approach before and like it.
Idea about W3C UI is very seducing.
Personnally, I'm needing framework without complexity of SUIR (great for professionnal using) cause I'm doing stuff as a hobby (but it's not incompatible with love good job with nice software tool).
I hope so @jlukic communicate about it and why not explain plan. It seem motivated people here who could help him to realize. Without that, SUI dead slowly...
EDIT : perhaps fork may be a solution but I agree people who wanna waiting more for @jlukic reaction about this
I spent my holiday getting out 2.3.0, you can view the full notes here, flexbox modals, font awesome 5, some other stuff.
I think the current speed of releases is healthy, although its clearly not the same as having dedicated full time work on SUI.
You can watch the Semantic UI discussion here.
The outline of general project modifications were
-- Setup CSS Regression testing
-- Setup Stalebot https://github.com/probot/stale to remove old conversations
-- Add new issue template
-- Consider bug hunt day
And for user contributions
-- Have system for code review for trusted contributors for PRs
-- Have volunteers to help maintain the discussion boards
Thank you for the new version and the information. This sounds really good to me, if we can work on these points I don't see a reason to fork.
Nice video !
I understand @jlukic point of view with core and modules as calendar, slider...
But perhaps community could developp "independant" modules working with SUI. So if you're needing image slider or other, you could find and use it. So core can keep slim for summer ^^
Seriously, I continue thinking community components is real advantage for dev : no need using obscure 3rd party components, no nightmare for maintenance.
@levithomason @jlukic What do you think about the opportunity to implement semantic-ui using styled components (for 3.0 maybe)? IMO this way we can merge sui and suir together and focus on maintaining one repo. With styled components we don't even need to use less, js have better data structures (simple obj, for example LESS doesn't have a nesting like objects do in js, so many vars have repeated words in their names) and styled components have a good solution for theming . You don't need to separate view and functional and can use fully-functional components pretty easily from the box. Styled components can be easily rendered on server & used on mobile with react-native.
Don't get me wrong, I just asking for your thoughts in this directions as maintainers.
I agree 99.9% on all points. I've written about this for a couple years now, as in Semantic-Org/Semantic-UI-React#1009. It would also enable sharing the same styles between the web (jQuery, React, etc), native mobile apps (React Native), and desktop (electron) apps. I've spoken all too much about the benefits in many issues over the years. I also have prototype of this on my machine now. Another contributor.
Jack and I have different opinions on this and he has good counterpoints as well. I still believe CSS in JS is not only the best solution but the future direction of the web.
The 0.1% difference I have with your comment would be the implementation library. I wouldn't use styled components, probably fela or some other.
As I've eluded to recently in a few issues, I may have a fulltime opportunity to make this happen in 2018. Really hoping it pans out.
For the uninitiated, you can see how CSS in JS has taken off by reviewing repos such as:
https://github.com/MicheleBertoli/css-in-js
https://github.com/tuchk4/awesome-css-in-js
Companies like Khan Academy have shipped open source libraries and have run CSS in JS in production for a while now. It solves, so, so many of the issues related to CSS.
Hello all!
What about crowdsourcing? SUI owners could create Trello board or something like this. They create small tasks for SUI community, and folks try to do these tasks. My idea is that owners have some free time, but their time isn't enough to write code. But their time may be enough to organize development.
I'm a mechanical engineer. There is a technical director on my firm, and he know what machine should be developed. He give me small tasks, so that I know what exactly and how I should to do. I don't know all moments about my machine. Every time I do only not big tasks, not more.
In my opinion tasks must be small. And person, who want and can help SUI, can do such task in one or a few days. As for SUI owners it should be easy for them to check the result of tasks made. This check should not spent much their time, and it should be easier for them to find free time.
Having css and js and html intermingled means that I'm forced to render with javascript. this means that either I have to render client side which means that there will be delay. If I want to render this serverside then I'm forced to also install nodejs along with the technology I use for my backend. So That a big 👎 👎
Also this means that I have problems to fill data with my backend technology stack. As the data must cross language boundaries.