JakeStaines

Yes, I know I can do that. The point is that it's a pain in the neck, and introduces more friction into the workflow. I switch mouse in a second or two - they're both on my desk and plugged in at the same time - whereas it takes much longer to open a menu, open a settings popup, select the right category, open a file browser, select the appropriate file, OK it, and OK the settings popup. And that also requires me to remember that I have to do it, which is non-trivial since Affinity isn't the only software I use and literally everything else I use supports both sets of copy/paste shortcuts and has done for decades. So what actually happens is that I work for a while after switching mouse, go to do a copy/paste, it doesn't work, I assume that I just missed the copy so I repeat, it still doesn't work, then I remember to go and change the shortcuts again. I tried this first, and if anything it makes things worse - I'm better off just taking my hand off the mouse and using the right-handed shortcuts. I'm sure you posted with good intentions, but please understand that it's actually quite frustrating to post a comment about/suggestion for an accessibility improvement and have a response within minutes telling you to do some cumbersome workaround you already know how to do that doesn't really solve the problem. I'd hope nobody would ask why wheelchair users want a ramp at the entrance when they could just go around to the back of the building, in the garage, up the car ramp, and through the service corridor, but that's frequently the kind of response one gets talking about software accessibility! It's worth bearing in mind that the study in the question at the top of the page in that link was conducted in the 80s, when most software was driven by heavily-nested keyboard menus and mice just didn't exist on a significant proportion of computers. And also by a company who had just released a computer with a relatively-novel mouse-driven GUI that was the USP of their main product. The more-recent (and given the degree of change in GUIs over the course of the 90s, more-relevant) studies in the top answer tend to suggest that keyboard shortcuts are quicker for simple operations by practiced users (single-shortcut things like change-tool, copy/paste etc.) and on a par with toolbars for nested commands for practiced users; the mouse is quicker for more-complex commands and for new users. I don't think it's really the case that most people think that nested sequences of menus and commands are faster with a keyboard. (And obviously this is all going to be very dependent on not just practice but also mode of use. Keyboard shortcuts will have an advantage in software where the users' hands are already on the keyboard all the time, because switching to a different peripheral is also a time cost. Mouse use will have a clear advantage in software the user is unfamiliar with as it has a far greater degree of discoverability. And mouse speed will have a lower ceiling for a lot of users as it requires coordinating hands and eyes to use effectively, whereas a keyboard can be operated entirely from touch.

Seconding this one - "like in videogames" and also like in most other software! This is as much an accessibility thing as a convenient feature, though. I'm old enough to predate readily-available ergonomic peripherals, so I have a bit of RSI that I've been fending off for years. I now use an ergonomic keyboard which has pretty much eliminated the typing stress, but mouse use - even with nice vertical mice - can be a problem after a while. So periodically, I switch mice - I have a left-handed one and a right-handed one, which means that I also switch from the right-handed Ctrl-C, Ctrl-V copy/paste to the left-handed Ctrl-Insert, Ctrl-Shift-Insert copy/paste. Affinity software is the only software I use where the left-handed shortcuts aren't pre-configured as an option, and on top of that it's impossible to configure them without replacing the right-handed shortcuts... which I also use half the time. I was convinced something was just broken when I first came across the issue; it seemed like something was failing to enter the clipboard or something, it didn't occur to me that the left-handed controls just might not be supported.

Cheers - sounds like a shame, I guess I'll carry on with V1 for now, then. As it goes, though - regarding #1, do you have any idea what that reasoning is? So far as I can tell it just seems to get attached to a particular set of appearance attributes at random and keep using them until it happens to change its mind and get stuck on a different set at some point in the future; if there's an actual deterministic process that I could work with - regardless of whether I think it's sensible - that would be an improvement! Sure, but since there's a sale on right now to encourage people to upgrade, I'd rather find out sooner than later! I'll keep an eye on stuff in the latest versions anyway - who knows, maybe there'll be some compelling new feature - but if it were the case that some or all of these issues were already resolved, I'd jump on the suite upgrade straight away.

I hope this is satire, because that's all fiction. All Neuralink has managed is to torture a load of monkeys to death, Musk has a long history of just making stuff up to suit his business interests. While the field of machine learning is a very interesting and fast-moving one I don't think any serious researcher really thinks we're remotely close to classical sentience. This is why the name "AI" is fairly crass and childish - machine learning models don't think at the moment, they're not intelligent. They're statistical systems that do a large number of calculations with pre-prepared "learned" probability data to deterministically transform a set of input data to a set of output data. It's an under-representation of their complexity and power to compare them directly to a photoeditor filter, but they "think" about as much as one. (That "pre-prepared probability data" is the bit that's a current ethical discussion around art - the big-name image-generation ML projects have all been trained on large collections of artworks harvested off of sites like ArtStation or DeviantArt without even notifying the artists who made them, let alone asking permission, and in some cases to the degree that the ML model can spit out a convincing facsimile if prompted to.) I don't doubt that if humanity manages to not kill ourselves off for long enough we'll eventually attain sentient, sapient, intelligent AI - the human brain is just a big chemical-electrical system, after all - but "the singularity" doesn't just require that, it also requires sentient AI that's intelligent and capable enough to design a new generation of sentient AI that's even more intelligent and capable than itself, which would hypothetically lead to a rapid exponential increase in AI capability. And there's a huge gap between "sentient" - something we can say of cats and dogs - and "intelligent enough to design and implement a new mind more intelligent than itself", which is something we can't even yet say of humans.

A couple of years ago I started using the Affinity suite to replace my ancient copies of Photoshop, Illustrator and InDesign, and while they're generally great for the price and for the most part I understand why feature focus has been where it is (as much as I'd like grep styles and vector fills!) over the years I've noticed a number of annoying quirks that add up to a kind of mediocre experience overall. I'm sure none of these can be intentional so I assume they're defects, but they're mostly inconsistent enough I've never found a reliable repro. Can anyone tell me whether these are things that they've also experienced in V1 but have been fixed in V2? I'm very happy to pay the upgrade price if my experience is going to improve, but I don't see anything compelling for my use-cases in the new features list so if the same quirks and annoyances are likely to persist I probably won't bother. The things that happen the most frequently are the following - although there's a myriad other things that are more intermittent so I only remember when they start happening again! These are mostly just for Publisher and Designer, which I use far more often than Photo. I never have any idea where the style used comes from when I start drawing a shape with the pen tool in Designer. It's not my current settings, it's not the style I have selected in the style palette, it's not the style of the last object I had selected (which is what I was used to in Illustrator) - and worse, it's almost always some obnoxious combination of characteristics that I set up for a one-off use and don't actually want ever again. Right now, I'm trying to do some small no-stroke flat-fill shapes to shade and highlight some lineart and whenever I start a new shape it keeps my current style for the first node, but as soon as I click a second node it abruptly switches to no-fill, a thick stroke, and giant circles on both ends of the line. This alone has had me on the verge of reinstalling Illustrator pretty much every time I use Designer. Leaving the app open for more than a couple of hours has it take up more and more memory and get choppier and choppier. I've noticed this with both Publisher and Designer. Closing it down and re-opening generally clears the problem, and I appreciate if I left it open for days I'm asking for problems like this, but it doesn't feel like I should have to do this when I've just been sat here working for a couple of hours straight. Printing from directly within Designer or Illustrator messes up a lot of vector shapes intermittently. Very commonly gradient-filled shapes have the gradient printed as an unclipped rectangle extending past their edges. Sometimes text jumps around and rescales itself, and in one print I did yesterday the fill of some text was squished and moved to the side, while the outlines of the same text stayed in the right position and proportion! Exporting the same document to PDF and then printing the PDF works fine, but I don't print often enough to remember every time and again it feels like I shouldn't have to. Clicking on the entries in the Appearance panel to change the stroke or fill settings seems to take a random number of clicks to open the settings pull-out. Sometimes I click the coloured circle next to a fill once and it opens up the settings fine; other times I have to click two times, and a third click closes it again; sometimes it just never opens, and I have to switch to a different selection and back again before I'm allowed to open the fill settings! Are any of these known to have been fixed in V2?

I don't think there's any point responding to such questions with anything other than a simple block and move on, frankly. It's not "most people" who make such assumptions, it's people with the special interest in machine-learning image-generation, that's all. Most people still see a nice picture and assume an artist made it on purpose. (I dislike the term "AI art", though, since really it's neither. Machine-learning algorithms produce entirely by rote, there's zero "intelligence" involved; and there's no intention, no creativity, so it's hard to call it "art" either. It's a fascinating and impressive technology which undoubtedly has a lot of promise for assistive workflows, but equally a lot of difficult ethical questions that so far have only been glossed over by its adherents.)

It doesn't matter what a Faction or a Program are: the point is that there are many different combinations of layers, several of which are shared across different output documents, several of which are recoloured via layer adjustments, and need to be exported in a number of known permutations. The point of the workflow is to make it as easy as possible to make changes in response to feedback without having to go and edit a load of different layers or perform a load of manual skill-less tasks that could in theory be automated. Here's a simple example file (attached): I've set up fake layer comps using Snapshots, so you'll need to open the Snapshots palette and restore each of those. Snapshots don't actually work for this at all as they record the state of the document at the point the snapshot was taken, so any changes made to the original layers won't get reflected - the point of layer comps the way they're implemented in e.g. Photoshop is that one could - say - make a change to one of the faction logos and just hit the "export my layer comps" button and get all the card backgrounds re-exported and ready to go, without having to do it all manually. If one needs to, for example, change the way the credits bar (second from bottom layer) looks - say rounding off the end, or pulling it across to fill the whole width - then one only has to do it once on the "Credits Bar" layer and then re-export layer comps, rather than do the change, and fiddle around with layer visibilities again to export nine different permutations of the file. If there's a way to achieve that - automated selection of pre-determined permutations of layers/groups, that update as you make changes to the document, and preferably are also selectable as separate exports by the Export Persona - then please do let me know. But I haven't found it and it seems to me it doesn't exist. layer comp example.afphoto

I don't think these details are at all relevant to the feature request, but as it goes: I'm talking about everything except the bird image and the text on the card. A very common practice when setting up cards for print is to start with an image or set of images which is the entire card background but without card-specific illustrations or text; then load those images into something like Affinity Publisher or InDesign and apply the game text and specific images for individual cards (the bird and the eye in the examples you posted) via a data-merge or similar. In this example case I would expect to produce a series of around twenty images* for all the different permutations of card type and faction - one image for use with all Shaper (green faction) Programs, one image for use with all Anarch (red faction) Programs, etc - and then hand those over to whoever was doing layout to do the work of merging all the game details into them. * Pedantic details: on one hand, in this specific case there's actually nineteen different card backgrounds required due to the details of how the game works, but that's irrelevant to the workflow problem; on the other hand, there may also be mask images to use to determine exactly which bits of the card-specific image (the bird etc.) get used, if for whatever reason the layout person is positioning those images over the top of the background instead of me leaving a transparent window and them positioning the card-specific image 'behind' the main card BG in their layer stack.

Sorry, but keeping files from getting corrupted is trivial. Leaving aside that I've literally never had it happen and I have not found the Affinity suite to be that unstable, backups, version-control software, even something as basic as Dropbox can deal with that - it's not that big a concern. What is a big concern is saving time; if one is working on something professionally then the ten or fifteen minutes this costs is a monetary cost, and one only has to perform such repetitive tasks a handful of times a month before - to be blunt - the Photoshop or Illustrator subscription pays for itself. And if one is working on hobby projects in one's spare time: speaking as someone with a full-time job and a family, if I can pay a relatively small amount of money to save myself significant time then that's money well spent, because it means I can actually spend my hobby time doing my hobby rather than tedious, repetitive, and error-prone work. Software is more useful the more it facilitates your work. I'm talking about creating the artwork in Affinity, that's why I'm talking about doing the adjustments in Affinity; there is no 'before import'. The above example has many elements which are the same from faction to faction, just using different colours; the most efficient way I know of to do that work is to create one version of the artwork and use adjustment layers to change the colours as appropriate to match the other factions' requirements. Without something like adjustment layers then I have to maintain four separate versions of the shared-but-in-faction-colours elements, meaning that if the art director asks for a change to them, I have to make an identical change in four places instead of just once in one place. Not only does this take extra time, it also introduces more scope for error - people make silly mistakes all the time, and the more of these skill-less tasks you can automate the fewer silly mistakes you'll get. Again: the goal is to reach the same effect without wasting valuable time. For what it's worth I'm probably more familiar with Designer than Photo, and - for example - did these game tokens entirely in Designer. Game cards do actually relatively frequently get worked on in Photoshop rather than Illustrator, for a variety of reasons, but regardless: Designer also doesn't have any such feature, so far as I can see, so one would have exactly the same problem there. I would agree that this would be a useful feature for Designer as well! Also, as I stated in my original post: I used the example of game cards because it's easy to explain, but my wife is a freelance illustrator who frequently works on full-screen illustrations for videogames, which most certainly would not be feasible to do in Designer; Photo is the correct tool out of the Affinity suite for that. I showed her this thread earlier and she informed me that the most complex scenes she works on can have over 50 slight variations over what is fundamentally the same illustration; in Photoshop she could make a small change and then easily re-export all the variants she needs to give to her client; in Affinity she can't, and it would be an hour-plus-long task to get the same result. So she continues to use Photoshop, at present. To be blunt: maybe I'm reading it wrong - and I apologise if that's the case - but your post comes off as a fan's defensiveness, as if I'm criticising your favourite software and it needs to be countered. I'm not; I like the Affinity suite, I've paid for Photo, Designer and Publisher and I use them fairly regularly for a variety of work. This is a feature request to improve the software and make it more useful for a variety of tasks, not a casual comment and certainly not an attack. I appreciate the Affinity suite is kind of an 80/20 version of the Adobe pantheon and is priced accordingly, but I'd wager that this kind of feature is more useful to more people than - say - Snapshots, which apparently made the cut and would almost certainly have been a more difficult feature to code. I've stated it as clearly as I can and I've stated why it's useful; I have a distinct workflow that I know works for my use-case, my wife's use-case, and clearly the use-cases of a lot of other people requesting the same thing. If you have a better approach that doesn't take more time, then by all means let me know! But telling me that it's not a good workflow is clearly false because it works absolutely fine for a lot of people in other software; and telling me that it's not a good workflow without presenting any alternative whatsoever is even less useful.

I appreciate that there's been requests to add Photoshop-like layer comps to Affinity Photo before (and frankly Designer would also benefit!) but they typically just seem to say "can we have layer comps like in Photoshop?" without explaining why that's actually useful, so rather than just add a "me too!" to the bottom of one of those threads, I figured it would be useful to give a comprehensive example of how layer comps are used and why none of the options in Affinity Photo presently seems to meet the same requirement: Let's say we're doing the card background images for the game of Netrunner, which I pick because there's an online card DB that makes it easy to link to concrete examples. (To be clear, I did not do the work for these cards; I'm just using them as an example of a very real workflow that is - so far as I can tell - tremendously difficult in the Affinity suite compared to other, similar software.) The game has cards of several different types, e.g.: Programs | Events | Resources and each of those types can be in one of several different factions, e.g.: Anarch | Criminal | Shaper As you can see, there's some common elements (e.g. the play-cost disc in the top left), some common elements which are recoloured from faction to faction (e.g. the art credit and copyright bar along the bottom), some elements which are present for each card type but recoloured by faction (e.g. the general card frame for a program) and some elements which are specific to the particular kind of card (e.g. the grey strength oval in the bottom-left, which only appears for Programs or the faction-specific background designs). Were I to be working on the graphic design for these cards, I would set up a single file with all the card graphics on, in which I would turn on and off layers, adjustment layers and filter layers to export the specific card layout for each combination of card type and faction. So I'd have my cost disc in the top-left visible on all my different exports, the credits bar along the bottom visible all the time but behind the layer that has the colour adjustment in, which itself has four colour adjustments for the four different faction colours, and so on. I could then turn on and off some combination of layers to export each combination of faction and card type, then use these as the backgrounds in Affinity Publisher or some more-specialised software like nanDeck to merge a spreadsheet of game data in order to produce all of the game's cards. Given that Netrunner has four sets of faction colours and five different card types, that's twenty separate exports for all the permutations. (Actually more, because it's an asymmetric game, but the other side shares so few of these elements I'd use a different file.) If I do this entirely by hand not only is it going to be tedious but also massively error-prone, so I'd want to use software that lets me automate the export process. The obvious place to start is the Export persona, but so far as I can tell the Export persona just... doesn't let you do anything remotely like this at all. You set up separate file exports from the same source .afphoto file by setting up slices, but these don't remember which layers are visible at all so if you set up multiple slices for the whole card area and try and change layer visibility on a per-slice basis, you actually just get multiple identical files output. So the Export person can't do this common export task at all, it seems. Which is a huge shame, because what the 'Continuous' feature on the slices promises would be incredibly useful for this kind of thing - if we could go around in the Photo persona making corrections to the file and have the exports just magically keep up and reflect the latest changes in a staging area without having to even switch to the Export persona, that would be perfect. The next most obvious place to start is Snapshots, and this looks like it solves the problem because you can indeed set up presets of different layers to turn on and off, so I could save snapshots of Shaper-Program, Shaper-Resource, Shaper-Hardware, Anarch-Program, Anarch-Resource... and so on. But then it turns out that Snapshots also revert any changes made to the contents of those layers since the snapshot was taken, which makes them useless for this purpose already. (I'm actually unclear as to why anyone would ever want to do this - I guess it's a "I'm about to make a big change, let's snapshot" option for people without version-control software to prevent saving fifteen WIP copies of a file?) Even if Snapshots didn't revert all the interim changes as well, there's no obvious way to automate the export of Snapshots anyway. I'd have to individually load each one, manually export the file, choose the export parameters each time (in case I want to resize the image on export or something) and name the file each time. So then I think that maybe I could set up a macro to automate the selection of layers/groups and the export of the files, but the macro feature can't record "Export File" actions, so that's a no-go. Maybe I'm missing some wonderful feature that would make this task easy, but I've played around in the software for a while, tried several things, and searched for options online, and nothing seems to actually solve the problem. Please let me know if I'm wrong! If the art director wants me to make the credits bar at the bottom of each card a couple of millimetres wider and re-export the cards, that's ten seconds of work that then takes ten minutes to carefully export and double-check all the permutations of. I'm spending 5% of my time on my actual job and 95% of my time wishing I was using Photoshop while performing the tedious error-prone busywork of turning various layers on and off and exporting each of the files necessary. To be clear for anyone unfamiliar: in Photoshop, that ten minutes of tedious, error-prone work is performed instead in a single menu item - export layer comps to files - and Photoshop just trundles through all the layer comps you've set up and saves them all to the path you chose. That's why people want layer-comp-like functionality in the Affinity suite: avoiding all that potential error and wasted time. And while card game graphics are an easy-to-explain use case, there's many more. My wife, for example, does freelance illustration often for visual novel games, which frequently have full-screen illustrations depicting events in the story. Often her client will ask for the same illustration to have several variations, so she'll set up various layer groups for each of the elements in a scene - maybe two people are playing basketball and the illustrations show just before a shot is taken, during the shot, and after. The BG would be the same, each of the line-art, colour fill and so on layers would be grouped by character, etc. If her client comes back and asks "could you just change the colour of this guy's socks to yellow" she opens up the layer group for that character, opens the layer group within that for colour fills, the group within that for socks, and changes that one element on the three different layers that contain the colour for each of the three output illustrations and hits the 'export layer comps' menu option and sends the output back to her client. Moving to a layer-group per illustration wouldn't be feasible because then she'd be hunting for the socks in three different hierarchies of layers and possibly miss one and have to go back and repeat the process, and sometimes these illustrations aren't so simple as just three variations - there could be twenty different variants showing a character wearing or not wearing a hat and/or carrying different things or whatever depending on the client's requirements, and because the files will be loaded by software they have to have precise systematic names. The lack of layer comps and that workflow in Affinity Photo has led to her completely writing the software off for her use, despite liking many other things about it. To my mind the ideal solution would be something that has a UI like Snapshots, but that just controls the visibility of layers. I guess a 'live-update' option on snapshots might be suitable, but it seems like it could be risky. It ideally needs to be a one- or two-click action to restore the visibility, as with Snapshots. Then to complete the workflow, we'd need an option on an export slice to select the layer-comp to export for that slice rather than just using whatever the current layer visibility is - preferably also with an option to automatically create export slices from each of a document's layer comps.

I recorded a short video of a snapping candidate choice problem as I'm not sure how I could easily explain it in words. Essentially it seems that after a couple of snaps, curve objects are losing their ability to be chosen and/or choose snapping candidates. It's not especially clear until they overlap, and I appreciate that it looks at first like I must just not have the hexes in my default-6 snap candidates list until I select them to demonstrate they're curve objects and not special polygons, but all three hexagons shown here are definitely in the recent-objects-snapping-candidates list throughout the video, the purple outline is just not showing up very well on YouTube but was clearly visible at the time. What seems to have happened is that it's somehow lost the ability to snap between these objects. In the video I open the snapping menu to show the current options and click on the objects a few times to demonstrate that they are indeed curve objects and no longer special editable shapes, but they refuse to register each other as snap candidates no matter where I approach them from or how they align or overlap. I then create a couple of rectangles, which snap to each other with no trouble, and snap to the hexagons, and then from that point onwards I can snap the hexagons again. This didn't actually last much longer than the video, unfortunately, before the ability was lost again in exactly the same manner, but again selecting the rectangles and then selecting the hexagons again seems to restore snappability. Other issues I'm having with snapping that I'm not sure whether are defects or limitations: The "snap to object geometry" option just seems completely broken. Nothing I do seems to allow me to snap node-to-node or node-to-line, for example, which makes assembling an accurate hexagonal grid (as I was trying to do when I found the above problem) basically impossible. Or am I misunderstanding what 'snap to geometry' is supposed to do? It appears that snapping is dependent on zoom level, in that even when these hexagons are playing nice and snapping to each other, if I zoom out too much they stop doing so. Zooming in again restores the ability to snap. I'm talking the difference between 700% and 1000% zoom here, so it's not even like I'm whacking right the way out to see the page as a tiny postage-stamp in the middle of the screen. (Although these hexes here are about 5mm across so maybe that's a factor.) (This I know is a limitation, but the maximum of 20 previously-selected objects to snap to makes tasks like assembling a grid of objects pretty tedious - is there no way to raise that? I appreciate that it'd be a bit of a churn for snapping in left/centre/right alignment with potentially thousands of objects on the other side of the page, but it'd be nice if snapping would at least automatically consider overlapping objects or something.)