Review My Startup : AppCanvas - Web Development Framework for Non-Programmers.
URL : www.appcanvas.com
This is very alpha and quite far from launch but I think an early feedback and guidance from the HNers would go a long way in shaping my strategy from here.
13 comments
[ 2.3 ms ] story [ 42.7 ms ] threadMy only suggestions at this point are to highlight browser requirements before someone takes the time to register; even though the registration form is short.
Secondly, do we really have to use Firefox?
However, the finished PHP applications should run properly on all browsers.
The UI for adding elements to the page is not intuitive. Some little "add here" thingie shows up and when I click then it gets added. Took me a while to figure out, I expect drag and drop.
It's cool that one can build a blog/HN, to-do list, but from looking at the builder I don't know how. But the bigger problem is the concept that I should build HN or a to do list from some low-level pieces. I think for designers it looks reminiscent of dreamweaver, where you couldn't build all that much without coding. For others it's probably too many buttons.
I think you should start with a finished template, let's say of HN and edit that to make a blog where users vote on one writer's articles, or a to do list, where you just remove the up/down voting and leave the submission form. Or to make some enhanced HN, add up/down voting conditions when a user just clicks the link, add "label" votes where you can vote for discussion article creates, not just article, then add site sort by quality of discussions.
An app builder is only as flexible as it's presets and templates, but starting with a blank canvas gives the impression that it's some professional tool with a big learning curve. I think you should underline that it's not flexible but enables you to prototype variations on established paradigms easily, so display finished template first.
I agree that one should not build complex websites from low-level pieces and AppCanvas was designed to address this problem from the very beginning.
- There are page templates (right now not many) which you can base your starting page upon. - There is a "Widget Gallery" which has high level widgets. - You can create a custom widget by selecting something and then reusing it.
Thanks for helping me understand how important having templates/presets it.
Also, I added a page called "test" accidentally clicked home, and now the page is gone, but I can't create a new page called "test" and I can't close the dialog to just edit the page I had.
Oh...I just worked out how to add an element. Why do I need to click "add here"? Now I'm clicking it and nothing happens....
Oh...You have to open the properties from the tool menu? Weird.
I like what you're trying to do, but I feel like I'm using Photoshop (You even mention photoshop in your intro), but it doesn't work like Photoshop. If you're going to use controls analogous to other applications, then it really needs to work like those other applications. You have a fairly low tech interface that LOOKS like a high tech application GUI.
In general, considering you're presumably targeting non-coders, it's VERY complicated. In reality, what is the likelihood that someone who doesn't know CSS know's the difference between padding and margin? Or that someone who doesn't know databases is going to be able to link up pagination to a database? You have a "power" application aimed at non techies...but you still need to be a techie to use it.
Who is your target? What problem is it solving them?
- Want to realize their idea of creating a complete functional app (as complex as Hacker News or Twitter) in few hours.
I agree that at the moment the interface sucks, the tutorials are missing and discoverability of so many features in the builder is super low, and I really need to work a lot more on them than I thought. But, I am excited that you liked the concept and took the time to evaluate it further.
Non-coders want to put as little effort into development as possible. Otherwise they'd be, well--developers. Starting with a blank canvas and the ability to add little widgets seems like a rather common sense approach to this kind of thing right? It confuses the hell out of us. How do you turn this blank screen into an actual application.
Of course education and intuitive UI will help with all of this, but I think you could also benefit by having a better understanding of how non-coders think about software design. I believe it was patio11 who made the comment that visual programming's greatest downfall is that it simply addresses the syntax. The designer still needs to understand the logic and data structures driving the whole thing. For example, I wouldn't just intuitively know that when I sign into a service, I become an object of the 'User' class based on a series of logical tests. This stuff matters.
As I said before though, I actually am excited about the potential of this project. I think you're aware of the issues I raised, you just need to iterate(again, and again, and again!). My email is in my profile, please drop me a line whenever you need feedback from a potential user.
Any half decent GUI driven development is going to have to need WYSIWYG, and currently in your app nothing is.
Sorry if this sounds harsh and critical - i have hope for this app, but it needs to improve before it can compete with the likes of yahoo pipes or blogger's template builder.