Currently in private beta, developed by Fog Creek, comes a online IDE for web applications, with deployment, hosting and team collaboration built in.
It reminds me how Heroku started (in 2007, 2008) with an online IDE for (only?) Ruby on Rails applications. Later, Heroku stopped with offering the online IDE and moved to the current deploy-via-git-push-model. It will be interesting how this works out for HyperDev.
True that also others can register. Their mailing did mention "As newsletter subscribers, we want to give you exclusive early access and get your feedback" below a "HyperDev Private Beta" heading.
Personally, I don't think it will be Open Source, at least not soon, but who knows. I haven't seen an indication on the business model or long-term strategy - could be paid hosting for running applications with more resources (just like Heroku, OpenShift, Google App Engine and other PaaS providers), or paid add-ons for additional features (monitoring, analytics, logging, persistency, high-availability). Or perhaps it is an entry for their Kiln service, version control and source code hosting, but I'm just guessing here. Fog Creek is using different types of revenue models for their services.
It's definitely not in private beta, because I also just signed up. It's a bit confusing because you don't enter a password -- you just enter an email address. They then assume that you want to authenticate via github to use that email.
HyperDev is taking a simplified approach to the space of live, collaborative coding and deployment. There's no need to set up your environment, configure your build pipeline, memorize Git, or manually deploy updates.
We're hoping that it makes development much easier for developers everywhere to collaborate, share, and have their code running live. We focus on removing obstacles and providing quick feedback. When you make a change to your project, the result is live on the internet for all to see in <1.5s.
I've found this kind of rapid feedback and shared experience to be really valuable, and we've got improvements planned to make it even better in the future.
Some feedback on what I'd like to see or couldn't find:
- preview option on the same page, C9 are the only ones I've seen that can do it but I think it's something many of us do with window managers, that is improved by being in the same app and tab especially on non-computers.
- Code formatters that run on save, most of the web IDEs I've tried only support manually running JS Beautify and they lack the ability to automate it or use any other formatters. Although it's part personal preference, some repositories enforce formatting before accepting pull requests.
- databases, otherwise it will often require a server or database hosted somewhere else plus additional software to access and build the database.
- Dropbox integration so you can always have a copy of your code close at hand and outside an IDE.
thanks for your feedback :) I'll try to respond summarily here:
- preview on the same page is being considered as an option on larger devices and on mobile/tablets where WM don't provide side-by-side capabilities
- Code formatting on save is a really nice idea :) We'll keep it into consideration
- We already provide an hosted object datastore backed by DynamoDB, check out datastore.js for the API and server.js for usage from this sample app: https://hyperdev.com/#!/project/bramble-leaf
- Integration with a source code repository and export to 3rd party data storage is being developed, it is one of our priorities :)
Seems to allow anyone to sign up... can you please explain how this differs from something like plnkr.io, codepen.io and JSBin, to name a few similar applications?
This is not only for frontend development and short lived code. You can write both front-end and backend code (check out the gallery for examples, https://cosmic-flower.hyperdev.space/ ) that is automatically and continuously deployed to a container and runs.
Just to clarify: some parts of HyperDev run in HyperDev.
EDIT: you can write your API, your little game for your son, your prototype app, alone or with friends, on HyperDev and it will always run, you can connect it to databases, use it as a webhook...
You should make this page more accessible from the site's root. The only reason I even clicked on the comments (and thusly saw your link to /about/) was to complain about the fact the HN link goes to a login/data harvesting^W^Wsignup page.
The link was posted by someone external, and it looks like it was only distributed by Fog Creek in an email with lots of information. I don't think this was quite intended for public consumption.
The link posted was to the main site / entrance point of HyperDev, and sure, it could have contained somewhat more information. I also found the link to pages with more background information like https://hyperdev.com/about/ after searching for a while.
Yes, we have plans to support renaming soon and custom domains eventually.
At present, you could use CloudFlare or CloudFront as a way to host your HyperDev project on your own custom domain, but we don't have any special built in support right now.
The design of this whole thing caught my eye, it's refreshingly unpolished and utilitarian but with a great deal of finesse and an appropriate amount of humour. It feels like a very honest, natively "webby" graphic design language. Lovely!
I would've liked testing this, if the signup prompts weren't so... aggressively displayed ? I almost never do signups, especially social signups via GitHub/Twitter/etc anymore for beta products because it's happened in the past that they ended up semi-abandoned with no way of deleting my profile/account, as well as for other reasons.
The alternative is of course, signing up with a temp email address and data, but it makes me feel uneasy going this route.
I suggest considering to make hyperdev.com/about the homepage instead (perhaps that's already the plan, once out of beta) or adding a link to it from the current homepage. People can read a bit about the product first, before deciding to signup, it seems more honest and empowering.
At least maybe display prominently on the homepage that this is a FogCreek product, I think it would make it easier for users to trust.
Try imagining a homepage visit is your first encounter with the product, having known nothing about it previously. It asks for your data, but it doesn't even tell you who made it. You have N other sites to visit or work to tend to.
sorry to hear that :( We are aware of this and completely understand it. Nevertheless, we preferred to take this route in order to test HyperDev in a "controlled" way. The experience we plan to have short after this testing period will be way more polished far less "email hungry" :)
First, this is awesome! I assume the reference to HyperCard is deliberate? (Aesthetically very similar too!)
Disclaimer: We're doing something similar at https://anvil.works (live web-app builder, though ours is drag'n'drop), so I ask from experience:
1. I see your sample app uses in-memory state on the back end. This is jaw-droppingly easy for simple examples, but dies whenever the Node instance "goes to sleep". Do you have plans for more durable storage? (I assume so, given that you're self-hosting - what do you use for your own storage?)
2. How do you sandbox user code? I'm going to guess "docker" but there's only so much poking around I'm going to do uninvited. I'm also guessing that allowing random users to run code (with outbound internet access?) on your servers is something you're going to have to be nice and careful with...
Hi meredydd! Thanks for your interest and yes, you got the reference to HyperCard ;)
1. Checkout the sample app https://hyperdev.com/#!/project/bramble-leaf for an integrated object datastore backed by DynamoDB, that works quite well! I myself built some nice stuff with it :)
2. We use a smart backend system to control user access, quota and sandboxing. For now we rely on honor system during the beta. We have systems in place to block abuse but... hopefully they won't be necessary :)
Neat - offering everyone a little bit of AWS quota is a nice starter. (We use Google Sheets, which has...different tradeoffs.) Is there a repository of such things that I've missed?
I also really like the way you've implemented the "your source is visible (presumably unless cough mumble monetisation?) but here are where your keys go" thing.
31 comments
[ 4.8 ms ] story [ 91.4 ms ] threadIt reminds me how Heroku started (in 2007, 2008) with an online IDE for (only?) Ruby on Rails applications. Later, Heroku stopped with offering the online IDE and moved to the current deploy-via-git-push-model. It will be interesting how this works out for HyperDev.
It looks lovely and streamlined for nodejs stuff, do you know if this will be open source?
Personally, I don't think it will be Open Source, at least not soon, but who knows. I haven't seen an indication on the business model or long-term strategy - could be paid hosting for running applications with more resources (just like Heroku, OpenShift, Google App Engine and other PaaS providers), or paid add-ons for additional features (monitoring, analytics, logging, persistency, high-availability). Or perhaps it is an entry for their Kiln service, version control and source code hosting, but I'm just guessing here. Fog Creek is using different types of revenue models for their services.
Since Heroku's famous pivot, there have been a few companies that have built online web (and other) development collaborative development environments, including https://www.nitrous.io/, https://codio.com/, https://c9.io/, https://codenvy.com/, and https://koding.com/. There's also the more scientifically oriented environments like https://www.pythonanywhere.com/, https://wakari.io/, and https://cloud.sagemath.com. I don't think any have had breakout success yet.
Of the above, at least https://c9.io/ and https://cloud.sagemath.com are open source.
Disclaimer: I wrote https://cloud.sagemath.com, which is why I watch this space.
We're hoping that it makes development much easier for developers everywhere to collaborate, share, and have their code running live. We focus on removing obstacles and providing quick feedback. When you make a change to your project, the result is live on the internet for all to see in <1.5s.
I've found this kind of rapid feedback and shared experience to be really valuable, and we've got improvements planned to make it even better in the future.
We've got some controls in place to lock it down if it starts getting away from us. Maybe "open house" would be a better way to phrase it.
I'll be here for a while trying to answer any questions people have.
- Source: I'm one of the creators of HyperDev
- preview option on the same page, C9 are the only ones I've seen that can do it but I think it's something many of us do with window managers, that is improved by being in the same app and tab especially on non-computers.
- Code formatters that run on save, most of the web IDEs I've tried only support manually running JS Beautify and they lack the ability to automate it or use any other formatters. Although it's part personal preference, some repositories enforce formatting before accepting pull requests.
- databases, otherwise it will often require a server or database hosted somewhere else plus additional software to access and build the database.
- Dropbox integration so you can always have a copy of your code close at hand and outside an IDE.
thanks for your feedback :) I'll try to respond summarily here: - preview on the same page is being considered as an option on larger devices and on mobile/tablets where WM don't provide side-by-side capabilities - Code formatting on save is a really nice idea :) We'll keep it into consideration - We already provide an hosted object datastore backed by DynamoDB, check out datastore.js for the API and server.js for usage from this sample app: https://hyperdev.com/#!/project/bramble-leaf - Integration with a source code repository and export to 3rd party data storage is being developed, it is one of our priorities :)
- A developer of the HyperDev Team
Just to clarify: some parts of HyperDev run in HyperDev.
EDIT: you can write your API, your little game for your son, your prototype app, alone or with friends, on HyperDev and it will always run, you can connect it to databases, use it as a webhook...
Here's some helpful info https://hyperdev.com/about/ to give you an idea what it's about. I'll try and respond to comments as I can.
- A humble HyperDev developer
Custom project names and custom urls are definetly in the pipeline.
- I'm the HyperDev designer
At present, you could use CloudFlare or CloudFront as a way to host your HyperDev project on your own custom domain, but we don't have any special built in support right now.
- the HyperDev designer
The alternative is of course, signing up with a temp email address and data, but it makes me feel uneasy going this route.
I suggest considering to make hyperdev.com/about the homepage instead (perhaps that's already the plan, once out of beta) or adding a link to it from the current homepage. People can read a bit about the product first, before deciding to signup, it seems more honest and empowering.
At least maybe display prominently on the homepage that this is a FogCreek product, I think it would make it easier for users to trust.
Try imagining a homepage visit is your first encounter with the product, having known nothing about it previously. It asks for your data, but it doesn't even tell you who made it. You have N other sites to visit or work to tend to.
sorry to hear that :( We are aware of this and completely understand it. Nevertheless, we preferred to take this route in order to test HyperDev in a "controlled" way. The experience we plan to have short after this testing period will be way more polished far less "email hungry" :)
We hope to see your feedback soon :)
- An HyperDev Developer
Disclaimer: We're doing something similar at https://anvil.works (live web-app builder, though ours is drag'n'drop), so I ask from experience:
1. I see your sample app uses in-memory state on the back end. This is jaw-droppingly easy for simple examples, but dies whenever the Node instance "goes to sleep". Do you have plans for more durable storage? (I assume so, given that you're self-hosting - what do you use for your own storage?)
2. How do you sandbox user code? I'm going to guess "docker" but there's only so much poking around I'm going to do uninvited. I'm also guessing that allowing random users to run code (with outbound internet access?) on your servers is something you're going to have to be nice and careful with...
1. Checkout the sample app https://hyperdev.com/#!/project/bramble-leaf for an integrated object datastore backed by DynamoDB, that works quite well! I myself built some nice stuff with it :)
2. We use a smart backend system to control user access, quota and sandboxing. For now we rely on honor system during the beta. We have systems in place to block abuse but... hopefully they won't be necessary :)
I also really like the way you've implemented the "your source is visible (presumably unless cough mumble monetisation?) but here are where your keys go" thing.