[HN Gopher] Show HN: FlakeUI
       ___________________________________________________________________
        
       Show HN: FlakeUI
        
       Author : tearflake
       Score  : 132 points
       Date   : 2025-03-03 05:29 UTC (1 days ago)
        
 (HTM) web link (github.com)
 (TXT) w3m dump (github.com)
        
       | alwa wrote:
       | It reminds me of the glory days when "hypertext" was a term
       | uttered with a straight face to great stroking of beards--
       | HyperCard, exercises in nonlinear narrative, VRML-based
       | "navigation," Apple eWorld [0] and the like.
       | 
       | > _Would you like to bring a touch of adventurous spirit to your
       | contents?_
       | 
       | I personally would not, but I'm really glad people more
       | adventurous than I are still exploring the periphery of UI
       | design!
       | 
       | [0] https://www.macworld.com/article/223467/remembering-
       | eworld-a...
        
         | freeamz wrote:
         | Or that Apple space based file system back in the 80's.
         | 
         | Try this a bit, it would be nice to be able to go directly to
         | the grand-child, instead having to bring up the parent before
         | going the child. Other wise can be a much better file naviation
         | system then what we have. Especially on touch screen I would
         | image.
        
       | drops wrote:
       | a brilliant idea in the correct direction of naturally-organic
       | UI, but the example site is rather slow in Chrome on an M3 Air
        
         | freeamz wrote:
         | Runs pretty ok on Chromium (degoogled) on Linux with 8GB ram
         | spec.
        
       | anorak27 wrote:
       | Reminds me of prezi[0]. It would be great if there is an open
       | source version of prezi similar to reveal js.
       | 
       | [0] https://prezi.com/p/p6evz0gdy5dr/ux-design-tips-for-
       | product-...
        
         | panglesd wrote:
         | Maybe you are interested in slipshow:
         | https://github.com/panglesd/slipshow/
        
       | unalarmed wrote:
       | I'd like to suggest adding support for clicking and tapping for
       | navigation. Having to drag feels unintuitive.
        
         | chipaca wrote:
         | Thank you for the comment. I would not have understood "can be
         | navigated using mouse" to mean "dragging".
         | 
         | Also I hate that I can't select text on this. Probably because
         | "dragging".
        
           | miningape wrote:
           | Exactly, clicking should be the default so the drag handler
           | doesn't prevent users from highlighting text - I literally
           | cannot read anymore without frantically double-
           | clicking/dragging on the words in the text
        
         | paxcoder wrote:
         | I think it would be pretty nice to be able to use the keyboard
         | to navigate the UI (left/right arrow keys to "orbit the
         | content", up/down to "zoom").
         | 
         | Using a pointer, I'd prefer to just be able to click on the
         | oval to zoom into, maybe double-click to zoom back out a level,
         | use the back button to go to the place I came from.
        
       | mindcrime wrote:
       | I really like this. And conveniently, I am just now working on
       | creating a new personal website[1] + blog, and I could very well
       | see using this for at least part of the site I'm building.
       | 
       | The only nit that I really have is that my intuition was that I'd
       | be able to select new "sections" (or "bubbles" or whatever
       | they're called) by clicking or double clicking. Having to grab
       | and drag isn't _bad_ but it violated the  "principle of least
       | surprise" for me a little bit. But not exactly a big deal.
       | 
       | [1]: https://www.philliprhodes.name
        
       | TechDebtDevin wrote:
       | Nice!
        
       | threekindwords wrote:
       | are you aware of this prior implementation [0]? it's now defunct,
       | but may give you some ideas!
       | 
       | [0] http://www.spicynodes.org
        
         | cdaringe wrote:
         | fun, but annoying to use. clearly not made for mobile!
        
       | remon wrote:
       | There's a significant performance issue. There's no good reason
       | for a few ovals and texts to stutter on my system. May be worth
       | investigating.
        
         | tearflake wrote:
         | May I ask, what machine you are running it on? On my Celeron
         | (4GB RAM), things are OK-ish.
        
           | recursive wrote:
           | Intel Core i7, 48 GB, Firefox.
           | 
           | Decidedly not "OK".
        
             | tearflake wrote:
             | I'm sorry, I really wouldn't know what's happening there.
             | On my Linux, Chromium works the best, Firefox can pass,
             | Opera is a bit slower, and Epiphany (is that the name?)
             | chokes a fair bit.
             | 
             | On i3, Windows things seem fine - I tried Edge, Chrome, and
             | Firefox.
             | 
             | On Mac, I somehow managed to get it working based on Lambda
             | Test web interface feedback, but I wouldn't know the real
             | use performance.
        
             | tracker1 wrote:
             | i7 is meaningless... it could be a 4c/8t processor from
             | over a decade ago that's slower then buttered toast today,
             | or something much faster with big/little cores more
             | recently.
        
               | mvdtnz wrote:
               | I don't think it's very likely that his 48gb machine is
               | over a decade old.
        
               | tracker1 wrote:
               | Looks like it's an 8th Gen with mixed ram running single
               | channel mode. The point stands though, the brand segment
               | is meaningless over time.
               | 
               | It's 5-6yo just before Ryzen got competitive with Intel.
               | A current option should be noticeably faster.
        
               | recursive wrote:
               | I'm going to be honest. I don't understand Intel
               | processor names, like at all. I thought i7 would give
               | performance to a first order, but if that's not the case,
               | I don't know how to tell what's in this machine.
        
               | spartanatreyu wrote:
               | Saying i7 is basically the same as saying you've got a
               | car with 4 wheel drive.
               | 
               | It doesn't say if the car is from the 60s or from a few
               | years ago.
               | 
               | You don't want to know if it's i3/i5/i7/i9, that's
               | essentially useless.
               | 
               | What you want to know is which generation it is.
               | 
               | For example:
               | 
               | Looking at a 1st gen i7:
               | 
               | - The i7-920 released in 2008 which only has 4 cpu cores
               | and runs at 2.66ghz
               | 
               | Compared to a 14th gen i7:
               | 
               | - The i7-14700k from 2023 which has 20 cpu cores, 12 of
               | which can run at 4.3ghz and 8 of which can run at 5.6ghz.
               | 
               | --------
               | 
               | Basically the way the intel naming scheme has been is:
               | 
               | "<i-number>-<gen-number><inter-gen-number><optional-
               | letter>"
               | 
               | - i-number (i.e. i3, i5, i7, i9) means rough idea of
               | which features are included. For example: i3 is basic
               | functionality, i7 will mean hyperthreading (although
               | competition from AMD meant that this feature started
               | become standard on all cpus)
               | 
               | - gen-number: (i.e. 1st gen through to 14th gen) means
               | basically what version the chip is. Kind of like an
               | iPhone 1 vs an iPhone 14, big difference at the start
               | then only incremental differences about 2/3rds of the way
               | through.
               | 
               | - inter-gen-number: (e.g. 400, 770, 900, etc...) means
               | basically how good the chip in this generation is
               | compared to other chips in this generation. Kind of like
               | iPhone vs iPhone Pro, vs iPhone Pro Max.
               | 
               | - optional letter: (e.g. nothing, K, F, etc...) means
               | basically any other info you should know about this chip.
               | For example: "K" means the chip can be overclockable
               | (runs hotter and requires more electricity for better
               | performance), or "F" means the integrated graphics on
               | this chip doesn't work so it's getting sold for cheaper
               | and you'll need an external graphics card to see things,
               | etc...
               | 
               | So the i7-14700k would be:
               | 
               | - i-number = i7
               | 
               | - gen-number = 14
               | 
               | - inter-gen-number = 700
               | 
               | - optional letter = k
               | 
               | A 14th-gen i7, which will be more powerful than an
               | i5-14600, and a "K" on the end meaning it can be
               | overclocked so we'll charge a little more on this chip.
        
               | recursive wrote:
               | Alrighty. Well, I've got an i7-8850H @ 2.60GHz. I know
               | it's at least a few years old.
        
         | einpoklum wrote:
         | In the author's defense - this is probably a prototype, hacked
         | together without trying to optimize performance. You should not
         | judge the idea by its realization's performance at this point
         | in time.
        
           | tearflake wrote:
           | It is true that I plan a custom renderer with scripting
           | support, to replace the HTML version. From the tests I
           | performed by now, it can be smoother and faster than the
           | current version. But it is yet to be seen how well would a
           | substitution to HTML catch up with the current state of art
           | CSS+HTML+JS.
        
             | tearflake wrote:
             | At some moment I had a version with cached bitmaps that
             | simply flew fast and smooth. But, since HTML has some
             | serious issues with rendering to bitmap from js, I had to
             | pick the slower version with native real time HTML
             | rendering.
        
         | edu wrote:
         | Agree, it's slugish on my M1 Pro/16GB on both Safari and
         | Chrome.
        
       | kstrauser wrote:
       | Make it themeable like Gabocorp[0] and the world will beat a path
       | to your door.
       | 
       | [0]https://www.webdesignmuseum.org/gallery/gabocorp-1997
        
       | QRe wrote:
       | Kudos for building something new, fresh & exploring,
       | experimenting.
       | 
       | I don't see a scenario where this would be useful. It reminds me
       | of exploded-view drawing but I don't see this being useful for
       | textual content. Do you have an explicit use case? The example
       | page, to me, looks very cluttered, overwhelming and IMO
       | aesthetically unpleasing when reading on a mobile device.
        
         | tearflake wrote:
         | Maybe a note keeping app?
        
       | runoisenze wrote:
       | Nice work. I wonder how the experience would be in VR.
        
       | settsu wrote:
       | This reminds me of the time a few years ago when mind mapping
       | sites and apps exploded into popularity among the... "technorati"
       | and sort of slightly seep into the wider online awareness but
       | then seemingly, just as quickly, disappear into the background
       | noise of the internet (I'm terminally online to a degree,
       | especially when it comes to tech news--and have a pretty decent
       | general awareness of pop culture trends--and can't recall having
       | seen the topic referenced since the trend faded. But perhaps I'm
       | just not in the right circles?)
        
         | fellowniusmonk wrote:
         | People mistake a helpful "view" for a useful UI.
         | 
         | None of these mind map, zoom first interfaces actually help
         | with creating a global understanding.
         | 
         | People take an occasionally helpful "view" for navigating items
         | and then mistakenly believe it should be turned into an active
         | interface for creation and editing.
         | 
         | Graph/Mindmap views should only ever be a view and maybe a
         | linking layer for nested text lists, actively operating in
         | these interfaces is worse for global understanding and systems
         | thinking.
         | 
         | I suspect this is because mind maps don't actually map to how
         | our brain stores information.
         | 
         | Visual programming and even tools like KNIME work for stepwise
         | workflow creation but they are not a good UI for new thinking,
         | it's too much UI for novel idea generation and brainstorming,
         | these interfaces are also useful for quickly understanding a DB
         | structure.
         | 
         | That's why they never take off and remain a niche tool for the
         | small number of people who have brain structures that find them
         | useful or are willing to bend themselves to an arbitrary
         | interface.
        
       | lacoolj wrote:
       | Seems to be a navigational flaw here - you can't get to any of
       | the intermediate "sections" without manually going back one by
       | one (other than "Home" to get to the absolute top)
       | 
       | Otherwise, this is pretty cool and would be great for one-way
       | traversing (maybe a quiz/test would do well here)
        
       | einpoklum wrote:
       | I tried the example site. The way this notion is implemented
       | right now, I don't see a benefit. One can barely see anything in
       | the nodes other than the main one; nor can one distinguish them.
       | TBH, even a tree of notes with titles, with collapse/expand
       | controls, would probably have been more useful in emphasizing
       | what FlakeUI offers.
       | 
       | Verdict: thumb down.
       | 
       | But - perhaps the example is only partial and perhaps this will
       | develop into something more meaningful.
        
       | chris_pie wrote:
       | This is very similar to the idea of Zooming User Interfaces (ZUI)
       | https://en.m.wikipedia.org/wiki/Zooming_user_interface
        
       | waingake wrote:
       | Have you seen / did you create this series on youtube? The final
       | video lands on something very similar to this.
       | https://youtube.com/playlist?list=PLsfH1Ahi4SzE-QmrsrtyZubGm...
        
       | jamesrom wrote:
       | Kind of cool for a presentation deck. Sorely needs keyboard
       | navigation support though.
        
       | humanfromearth9 wrote:
       | New feature request: support for nix flakes...
        
       | com2kid wrote:
       | Super fun, but not being able to scroll text within an oval using
       | the scroll wheel kinda sucks on desktop.
       | 
       | The mobile UI is really slick though!
       | 
       | I feel like this UI would be good for some sort of narrative
       | game, where each oval is a room and things keep changing in each
       | room.
        
       | monkeydust wrote:
       | Liked this more than expected. Here's my possible use case.
       | 
       | In a work setting I am constantly juggling between
       | excel,PowerPoint,word and bunch of online wiki style tools
       | 
       | When I have to craft something that requires information from
       | these tools, like many, I end up creating a PowerPoint
       | presentation.
       | 
       | Gets the job done but never feels ideal given the one way
       | directional nature of it, yes you can hyperlink but gets messy
       | for audience.
       | 
       | This style of navigation could help especially where you are
       | layering the information across multiple themes (each theme say
       | is a node). As audience you might go deep into one node quickly
       | then zoom back out to understand how that node interacts with
       | another.
        
       | thih9 wrote:
       | In practice surprisingly similar to switching between apps and
       | tabs on modern ios + mobile Safari.
        
       | jeffhuys wrote:
       | Makes me think of arguman[1] and kialo[2]. I discovered arguman
       | maybe 10 years ago and have been thinking about it ever since.
       | It's non-functional now but the source code is available. Maybe
       | it's just my wishful thinking, but if we could map out every
       | detail of an argument, maybe we could make progress faster, as a
       | species. If this were incentivized well (like for instance
       | polymarket), maybe with crypto rewards or something, this *
       | _could*_ take off.
       | 
       | [1] https://github.com/arguman/arguman.org - the website exists
       | but is non-functional now
       | 
       | [2] https://www.kialo.com - feels like a dumbed-down version of
       | arguman
        
         | bflesch wrote:
         | interesting. I've been thinking about building something like
         | this to provide a structured framework for social media
         | discussions and argument chains.
        
         | keyserj wrote:
         | I think if we could improve our ability to easily & precisely
         | communicate arguments, that would be huge for humanity's
         | ability to deal with problems. I'm building a tool[1] that I'm
         | hoping can help with this, maybe be a successor (in some ways)
         | to something like Kialo.
         | 
         | It's critically missing suggestions/approvals[2], and perhaps a
         | simpler interface[3], but I think that, for arguments about
         | problems/solutions, the idea of grounding arguments within a
         | cause/effect diagram is really powerful both for getting on the
         | same page and for making concrete progress towards improving a
         | situation (rather than arguing for argument's sake).
         | 
         | I'd be happy to get your thoughts on it if you have any (I'll
         | be making a HN post about it sometime soon^TM).
         | 
         | [1] https://ameliorate.app/
         | 
         | [2] https://github.com/amelioro/ameliorate/issues/11
         | 
         | [3] https://github.com/amelioro/ameliorate/discussions/541
        
       | dvngnt_ wrote:
       | Maybe the arrow keys could have a different color if you can
       | actually navigate in that direction.
        
       | ramesh31 wrote:
       | Shades of ZigZag
        
       | sebastianmestre wrote:
       | Runs slow as hell on my phone. Around 10fps.
       | 
       | My phone is not great but drawing some circles and some text
       | should not take 100ms. On any device.
        
       ___________________________________________________________________
       (page generated 2025-03-04 23:02 UTC)