Peter Werner

AI itself is not the issue, but erosion of users' control over their content and devices. I think Adobe management has realized that generative neural networks allow them to reach a much bigger market by not only offering creativity tools, but also the creative services themselves. Not everyone possesses the necessary skills to create something that looks good and professional, but almost everyone needs creative services at some point in one form or another, and demand will be even higher when it becomes available at commodity prices. Templates were the first step in this development years ago. And the plan is likely to transition from serving the market of software for creative professionals to the even more lucrative automated creative services, with Adobe Express being their first product in this space. The flipside is that the creative professionals will first have their skill and content exploited (those who offered stock photos on Adobe Stock were first) and then be treated as an afterthought since the amount of money being made with them pales in comparison to the amount of cash in the new, much bigger target audience. This will also be a huge disadvantage for those clients who still require custom-tailored creative solutions, since they, too, will become a very specialized niche market. While this is all nice for someone who just wants a quick wedding invitation that looks good for cheap, this will have major negative effects: On the one hand, very few people will want to learn high-end creative skills anymore because it will require a lot of dedication and tolerance of frustration to first match the level of generative networks and later surpass them. I mean, why painstakingly learn and practice to do something manually for years on end when the AI can consistently give you superior results in an instant? On the other hand, it will lead to a lot less original creative creations by humans since fewer are needed, leading to a decline in training material for generative networks. Let's compare this with the online search market: If major players in the search market like Google and Bing replace their current search products with those occasionally reliable LLM chatbots like ChatGPT (terrible idea, in case there was ever any doubt), people will now get their answers from them directly and very few will visit the websites that provided the source for this information. As a result, the incentive to actually offer websites will decline steeply. That means availability of independent information will stagnate and there will be a concentration on a few major players like press agencies and publishing houses with contracts with the search companies, plus some social networks with their user generated content. However, I don't really see such institutionalized sources for fresh training data in the creative industry on a comparable scale, aside from the occasional photo agency, so maybe Adobe will continue offering their software, possibly even for free, as a way to constantly source new training material and users will have no choice but to be the product. Or maybe at some point the dataset will be considered "good enough" to consistently make money and they will just discontinue their professional software. And just like in the search and information space, I see a danger for monopolization in the creative space, too, since not a lot of other companies have access to a dataset as big as what Adobe will have access to since they control the tools people use and can turn them into a trojan horse. I hope Serif/Canva will not head down a similar path, but keep serving that smaller custom high-end market with products that respect them, their privacy and their work. Also, I hope I'm wrong in general.

Any feature that uploads full or partial content to a server without end-to-end encryption (end-to-end encryption in this context meaning the server can't access the data) should display a big fat red warning and a confirmation dialog informing the user of the full implications. I don't even want to think about how many creative professionals knowingly or unknowingly upload their customers' data to cloud services without their consent these days because software and platforms are designed to deliberately make this both extremely easy to do by accident and extremely difficult or impossible to work without. By the way, Adobe also upload your (hashed) contact information to Meta platforms for use in ads and also makes countless connections to a variety of servers, including anayltics services, in the background when their software is just installed. That means they know things like your IP address or even when you switch on or off your computer, even if you don't start any of their applications. Aside from the creative space, they are also a major player in the ad and tracking business. And you can't even use many of their mobile applications (eg. Adobe Fresco) without your content automatically being uploaded to their servers, whether you want it or not. I think that Serif resisting these trends is a major part of why people choose to migrate to Affinity (aside from it being good and affordable software of course), and considering Microsoft is moving in a similar direction with Windows and Office, and with restrictions Apple as well, I think Linux support would be a good logical next step in offering a true alternative for those looking to escape these highly concerning developments in the tech space.

I hope I haven't missed this being mentioned in this giant thread, but one thing that would be very useful is a way to associate custom metadata with an object. Say my plugin generates some type of shape based on parameters, say a barcode, a spirograph pattern, an extrusion or a mathematical function graph. If I want the user to be able to change these parameters after the fact, it would be useful if I could get access to the parameters that my plugin used to create the object in order to present them to the user again and re-create the object based on any changes the user made. Another use case would be to link the object to an external item using an ID. For example, let's say I make a plugin that offers to paste text content directly from a CMS like a WordPress blog. If I can store the post or item ID, I could offer to the user to refresh the content from the website when changes were made. Same for an editorial publishing system or a database. Or if I made a plugin to be able to place LaTeX formulas, the LaTeX code the user types into my UI could be processed to PDF using a command line tool to a particular asset folder and placed as a linked graphic. If that graphic had a field with the associated LaTeX code, I could then allow the user to edit the code and update the linked PDF from within Affinity. Another use case would be a script that creates diagrams based on a CSV file. If I can save the parameters used for that object (colour palette, axis and label settings etc.) and the path to the CSV file, the script could offer an "update" button to update the diagram when the CSV file changes and the user wouldn't have to re-do all their colour and axis settings. One more example: Say my script integrates ChatGPT into Publisher to generate text. In this case, I might want to associate the prompt that was used to generate the story and a hash value of the generated text with the text frame. That way the plugin could offer to edit the prompt and have the content re-generated, but using the hash it could determine that the text was changed and display a warning that manual changes would be lost. Links/References to other objects in a document would also be interesting. For example, if my script was to, say, generate a set of rectangles with the most prominent colors in an image, sort of as a colour palette, it would be useful to be able to reference the corresponding image object inside that document from the object group that was created. Similarly, if I made a script that does vectorization. In some cases where external files are involved, an integration with the Resource Manager might be interesting, but I would consider this optional.

Just ran into this in V2.5. I agree this would be a very useful feature inside the Font Manager.

It is entirely possible that this has changed, I have not conducted any tests myself recently. Also, different parts of Photoshop (such as Camera Raw) likely have their own separate way of representing image data.

That's why Adobe decided to actually use 15-bit for what they call a 16 bit mode inside Photoshop. That way they can have an exact midpoint representing 50% gray.

I think there would be some major potential for Publisher as well when it comes to copyfitting and tweaking composition. Such a feature could be used to help with making a text fill a frame for example, or to fill gaps between words that are visually too big. Say you need another line of text in your frame and you click a command that offers a few good alternatives to choose from for adding filler words, very minor rephrasing etc. to the story. The proposed text changes could be shown in some sort of diff view or in-place in the layout in a highlight color with the option to accept each suggestion into the story or reject it. Conversely, if a story is a few lines too long, it could offer suggestions for which parts to cut first. The part with the manual review of the suggestions is important because I don't think such a feature should just rewrite a text on its own without control, just speed up the existing manual workflow by offering options.

In very simple terms, you can think of Wine (and Darling for that matter) as something similar to Qt or Gtk+, combined with a tool that knows how to load a windows exe file and start it so the software developer doesn't have to recompile their software for Linux against the WINE libraries. So if a program says "Create a Window" to the Qt library, the Qt library would then call the underlying X11 command. Similarly, if the Windows program calls a Windows API function like CreateWindowEx under WINE, the WINE library would in turn talk to X11 and call the underlying X11 command to create a new window. So it's not inherently more inefficient than most major native Linux applications these days. Apple QuickTime, iTunes and Safari used a similar method for their Windows ports where they re-implemented Mac APIs on top of the Windows API. Carbon was pretty much the same thing for building and running OS 9 programs on OS X/Cocoa with only minor code changes and was used by major software like Adobe Photoshop. MS Office (and I believe Internet Explorer for Mac too) used to be based on a re-implementation of the Windows API for Mac (in fact, you can still find references to building on Mac in Windows API header file code today). Many big-name programs like Corel Painter, ProTools, Avid MediaComposer and others used a library called Mac2Win from a company called Altura Software to port their OS 9-era software to Windows. Corel used WINE libraries in a similar manner for their CorelDRAW and WordPerfect for Linux, as did Borland with their Kylix IDE. The office suite GoBe Productive re-implemented parts of the BeOS API for Windows to make their Windows port. So while the magic is a bit more visible when running WINE, a lot of successful and somewhat performance-critical programs have used this approach in the past.

It is likely that the reason why an increased number of people have been migrating from Windows/Mac to Linux or back to Windows 10 recently is the same as why a lot of people have moved from Adobe software to Affinity. Hence I wouldn't be surprised if there was a higher than average overlap between people who would like to use Linux and those who already use Affinity or are thinking about making the jump. In other words, a Linux version might be another incentive to get more of those Adobe users to switch, particular if an announcement is timed right with respect to the end of support for Windows 10. People in the graphics and media space have to choose their OS based on where their software runs. Only looking at the data of the status quo ignores the chicken and egg nature of the problem.

I'd suggest creating the specific design with border and text that you would like to reuse over and over again as an asset in the Assets panel, with a dummy QR code inside. Then instead of adding a QR code with the tool from the toolbar, you would just drag your asset into your document and edit the placeholder QR code. It wouldn't be slower to use than the QR code tool, but it would avoid complicating the tool with a hundred options for all the different border and text combinations people could possibly want to create. But one thing that might be nice would be an integration with the text variables in Publisher so the target URL of a QR code could be placed somewhere as plain text in a way that it auto updates when the QR code is modified. This is a common use case to cater to peolpe who have trouble scanning a QR code for whatever reason. For example, a "Copy Target URL as Text Variable" context menu command that would then allow to paste it into any text object in the document.

Thanks for the reply! For some reason the search function did not give me any resultswhen I searched for a workaround before posting, but now I do seem to get several results. Maybe I misspelt something or just used the wrong terms earlier. Sorry for the noise everyone!

The "Area" setting in the PDF Export dialog box reverts to "All Spreads" every single time the "Export" command is invoked on any document. Other settings in that dialog box seem to be more persistent, so I assume this is a bug. I have to use the "All Pages" setting quite often, hence this can become a bit frustrating to change every time. And if I miss it, I have to go back and re-export the PDF one more time. Affinity Publisher 2.4.2 on Windows 10.

I did some more experimenting. It turns out that renaming the assets.propcol file in C:\Users\[username]\.affinity\Common\2.0\user and having Designer create a blank one from scratch on next launch does not fix the issue. Therefore it's probably safe to assume that it's not assets.propcol getting corrupted either. It also seems to be irrelevant whether I drag a file from the Desktop or an Explorer window showing an SMB network location. Next thing I can think of is system language – my system is running Windows 10 Pro 22H2 with non-English locale.

@stokerg Thank you for looking into this! As far as I can tell, pretty much any SVG or JPEG from any source I tried triggers this bug, so I don't think this is related to the file being dragged into the panel. I even created an empty new Designer document, drew a rectangle inside it, saved it to the Desktop, closed the document and dragged the file icon of the .afdesign file from the Desktop into a subcategory in the Assets panel. Results are exactly the same. Let me know if you would still like me to upload a few random files that are known to have triggered the bug (though, again, I haven't found any that don't). Alternatively, I could send my assets.propcol file if you think that might be the cause. By the way, contrary to what @Return stated, toggling "Show as List" will not cause the new asset to appear for me. But using the "Sort by …" commands indeed does.

Thank you for your reply! The assets.propcol file is currently around 124 MB, so not giant considering that it's not uncommon for even just a simple icon pack to be up to 15 MB. In my case it's 6 smaller asset packs averaging maybe 5 MB each plus two larger ones I put together from an existing library of material, one 10 MB and one about 56 MB. Even without those two, the panel never felt really snappy. It is probably safe to assume that it's common for a heavy user of the Assets feature to be working with a library at least this large, if not significantly larger. I suspect that something time consuming like rebuilding all thumbnails or moving a lot of data on disk gets triggered every time an operation is performed in the panel.

When dragging files from Explorer into the Assets panel, the interface does not refresh until another action is performed that can trigger a UI update, such as reordering subcategories, dragging an element in from a document, deleting an existing asset or similar. Moreover, the panel can get slow, such as when dragging elements from one subcategory to another (which by the way can get tedious because there is no way to multi-select items) or deleting items. Even just dragging to reorder subcategories can cause a lag of one or more seconds after the mouse is released. Designer Version 2.4.2 on Windows 10

I just ran into similar problem with a placed CMYK AI file which would always get rasterized when exporting to PDF. I was able to get it to work by going to the options bar of the selection tool in Publisher with the object selected and under the "Layers:" dropdown choosing "Show All". Despite never having changed any of those layer options and all layers being on anyway, choosing "Show All" resolved the export problem. EDIT: I found this thread using search, this is referring to Version 2.3.1

Agreed. For me, what still keeps me away from using Publisher for more projects instead of InDesign is primarily the lack of an equivalent to Adobe's multi-line paragraph composer and Adobe's much superior Optical Margin Alignment feature (in Publisher, activating that option without extensive manual tweaking just makes everything worse unless the font has specific information embedded, which most fonts do not). I hate to say it, but especially text in narrow columns still looks signifantly better in InDesign. I'd be very happy if that changed.

I have to say, your process is pretty quick that way! Still, I think it would require someone to think about the problem for and extended amount of time or research it and then to add custom shortcuts for two commands to get there. Selecting might also be a bit slower in a complex real-life situation with more objects around, and the object would stay in the same position in the document hierarchy. Nevertheless I think this is the quickest solution we have so far. I'm sure something along the lines of the Find & Replace plugin you mention will pop up very quickly as soon as we have an API. But it seems more suited for automating more complex tasks. It looks to me like using that to quickly swap two objects would take a similar amount of time as doing it manually would. But then again I have never used that plugin.

Thank you for taking the time to clarify. My intuitive assumption would be that such a command would operate with respect to the effective bounding box, with all transforms except translation basically "baked". I think your post also serves to illustrate another reason why the naive approach of copy/pasting coordinates may not yield the desired result in all cases. Another question that arises is what should happen if the two objects are children of other objects, such as members of groups. In this case, I would assume that they also swap position in the document hierarchy.

The great thing is, purely mathematically speaking, the Lift/Gamma/Gain/Offset wheels don't do anything different from what is already possible with the Levels tool. It's just a different interface. That means if this feature is added to Photo, it could be written to PSD files as a Levels adjustment layer so the file would still come out looking correct in Photoshop or other software. DaVinci Resolve, 3D LUT Creator and similar software have tons of great tools that clearly show that Adobe's color adjustment toolset has not developed much since the 1990ies. Some minor advancements have made it into Camera Raw, but not a lot. I'd be very happy to see some more of these great tools from the film and video world make it into Affinity.

My suggestion would be to respect the setting of the transformation origin widget in the Transform panel to determine which anchor to use for position switching. With respect to individual object transform, I would suggest to just keep all local transformations like rotation/flip/skew etc. associated with the object since this would match what most people would intuitively assume the result of such an operation to be: I think anything more advanced than that would need so much user input that it would defeat the purpose as it would no longer be a one-click operation. That being said, I could indeed imagine a "Switch Attributes" command or script that gives the user a dialog box with checkboxes for granular control of which attributes of an object to transfer and whether to switch or copy attributes between the objects. So you could check or uncheck "position", "rotation" etc. individually, but also "effects", "style", "stroke attributes", "fill attributes", "blend mode" etc. Another approach might be a "Copy Attributes"/"Paste Attributes" type of set of commands. Similar to corresponding features in Adobe Lightroom (pictured below), DaVinci Resolve or Autodesk Maya. But again, that would solve a different problem.

My estimate would be that for an average serious user of Affinity Publisher, swapping the position of two objects is a task that comes up at least once or twice per session, whereas the use case you give is much more specialized. If I had to guess, I would estimate that they differ by at least a factor 1000 in frequency of relevance for the average user. If you say it's something that comes up in your work a lot, I'd be happy to include such an option for you if we end up having to go the scripting route. I am sorry if you got the impression that I somehow do not value your contribution to this discussion. My point was that even with the faster process you suggested, it's still significantly less efficient than a dedicated solution. It also has other disadvantages, such as the fact that the software only displays floating point numbers with a limited precision in the UI, so copy pasting coordinates may cause position shifts in some cases. I'm not sure how to respond to this in a constructive way – the reason why we are having this discussion in the first place is that I don't like that procedure.

I did write one for breaking a frame into columns in InDesign at some point. I ended up using it so often that I wondered why they never added it to the core feature set. Could you give some other examples? Most of the ones I can think of are already there ("Align Vertically", "Convert Art Text to Text Frame", "Swap Stroke and Fill Color", "Duplicate Layer", …) I'm actually genuinely curious because if these things aren't added, I'm thinking about writing a script as soon as we have support for that. I did. My point is that these are not one or two click solutions. 12 seconds versus 2 seconds easily adds up.

It's really interesting to see all these different ways to achive the desired result, but I think these demo videos being 20 seconds long are a testimony to the fact that there is indeed a potential to save time with a dedicated command. I think it's one of these small things that can really add up if you use a software every day. All of these can be achieved in some other fashion, but after a while of using these workarounds, it just feels like there should be a faster way. A similar thing that I run into constantly is whishing for a command to quickly break a multi-column text frame into individual linked frames (corresponding feature request post here).