It's built entirely from the ground up – no ui libraries. It's focused on simple, PowerPoint-like usability instead of Webflow's complex, Photoshop-like approach.
How does it scale to real-world projects? One of the problems I’ve felt with Webflow is that you’re missing all the tools you would normally have when making a website, like version control, being able to search and replace across all project files, having a decent deploy process, or being able to automate or refactor things without hundreds of manual actions. You also still kinda need to know HTML and CSS to not make a mess (I’ve been tasked with cleaning up a mess like that, and it’s not much fun, especially when the toolkit described above is not available). These are flaws that are hard to anticipate until you’re halfway through a project.
Considering you’re describing Aspect as easier, what are your solutions to these problems?
While I'm new to Webflow, I get a feeling that some people try to push it too far. It seems to me that it's mostly a no-code tool meant for basic "brochure" sites with some modest CMS capabilities built in. And for many, many businesses, that's sufficient. It provides additional design flexibility for those who understand HTML and CSS. If you're making a site that's so complex that you need version control, you should probably be using a web application framework. But, in a pinch, you can always export your code and do VC on that. Just my $.02.
Those are basically the types of projects I’ve been involved with, but with high standards for the design quality, and sticking to a brand manual. A project like that still has the kind of complexity where version control would be extremely useful. Like just having a place to see what changes have been made, by whom, merging the things we want to deploy, reverting other things - it’s just a matter of time before you run into problems without it. See also Elof’s comment in the thread which explains basically the same problems.
Great questions! I’ve got upcoming features for version control – it’ll still involve conflict resolution in git, but much easier than current code export methods. So far I’ve focused on good code export readability to address integration into existing codebases.