For a static site that one can drag-and-drop to, I'd just use S3's new 'website' bucket configuration with CyberDuck as the S3 client (folder sync or drag-and-drop files).
Oh, I wasn't going on about that specific client (I am also really happy with Transmit) just about the thought that people need to be warned about paid software.
I'm currently building a site generator for photoblogs hosting directly on S3. Seems like good performance, cheap hosting and no more worries about Wordpress
It's always being hacked/targeted, requires constant updating, needs plugins for some functionality, and these may throw a hissy fit when you perform one of the many essential updates. Basically it's a general tool and a victim of its own success.
I'm going to built something that fits my exact needs and produces a static site hosted on S3.
"... I am thinking of moving my blog there (statically generated jekyll site). Can anyone comment on whether it's practical or not? ..."
Practical yes, you can even use your own domain, cf "Custom Domains". Jekyll is a great bit of software, I've read through all the source to see how it works.
Why add dropbox into the mix? You're essentially editing files directly on the server, since every change gets propagated live whether you want it to or now. If that's what you want to do (and remember, you never should), why not just do it directly?
Dropbox means you don't have to edit files directly on the server, it means you can edit them directly on your computer. That's going to be simpler and easier for some people.
This service apparently runs a Markdown-to-HTML converter on your Dropbox subdirectory whenever you change a Markdown file.
What about client-side rendering of Markdown? You'd have an HTML file which includes a separate Markdown-to-HTML converting JavaScript, and an invisible pre containing the markdown. That HTML file, consisting mostly of Markdown text, can then be edited in a simple text editor.
I've taken to doing basically this with PHP for static websites. A lot of websites I work on are really just a handful of static pages--- needing logins and a database and a "theme" and to fret about security patches just for that is like swatting a fly with a bazooka.
It's a supporting site for a marriage seminar being put on by a local church---the content on that page is being sourced from http://8cities.org/content/newyork-left.txt and http://8cities.org/content/newyork-right.txt , so it's easy for the staff to make changes to the files without needing a ton of technical know-how, but there's less overhead than with a whole CMS.
Another way to do it that I've found works well is to have a php file with a little <?php ?> block at the top where you set template parameters, etc, and then have the markdown content following in chunks separated by a marker (like #######). Then you include the file in your processing code, use PHP's output buffering to capture the content, split it into chunks, and run markdown and smartypants on it. Then you have HTML blocks in $content[0], $content[1], etc, which can be included into the template PHP file.
An approach like this would work really well with a Dropbox backend.
So instead of rendering the markdown to html once on your computer, you're going to do it every single time on your users' computers? What possible benefit could there be for this, besides unnecessarily burning peoples CPUs?
You don't need to regenerate any HTML and changes are instant. Also, I honestly don't think one can burn through too much CPU with occasional JavaScript. This is some hardcore environmentalism.
It's not environmentalism, it's just common sense politeness. There is no reason a users web-browser should have to render something twice just for static content. All just to save you a single `make` before uploading.
I've been joking for a few years that it's only a matter of time until some joker implements a browser in javascript/html5 and the rest of the web development community enthusiastically adopts using it as standard practice. I'm starting to fear that this is more of a possibility than I initially thought.
I doubt that translating a few hundred lines of Markdown into HTML with an efficient JavaScript JIT is noticeably more expensive than, say, decoding a big JPEG picture or displaying a Flash ad, or having menu items slide in and out.
I think this is a great step in the right direction, publishing is getting simplified and with the advent of cheap storage that can be easily placed behind a webserver the options for static websites are quite interesting.
I foresee more services of this type and more power to you guys for building this, I highly recommend exploring and iterating around more simple ways for people to 'create webpages' by simply dropping 'something' 'somewhere.
I would like to record for posterity that I was coding up pretty much exactly the same thing over the weekend without knowing this already existed, so when I launch please go easy. I have commits as proof!
Actually this makes me wonder, why dont most web-views of VCSs dont come with a web based text editor? Would be ideal when working behind firewalls/proxies.
This quote from the article about dropbox..."Even though overpriced in many ways, it is simple and easy-to-use and save my asked [sic] a number of times."
I'm up to 11GB free, it's possible. I've sent them over 150 referrals, most of whom've completed signup. Plus doing things like taking the tutorial, etc.
The referral model by dropbox must be awesome for customer acquisition.
150 referrals, wow. Do you experience some kind of saturation by now, in that it becomes increasingly harder to find a potential new user who hasn't yet subscribed?
actually I don't put any work into getting referrals at all, so I don't know. I wrote a blog post about it and put the link on twitter a few times, and the referrals just keep pouring in - I'd love to know where they all come from to be honest! I've hit the limit as to my free space allowance now.
If you compare the price with solutions without the syncing experience, you can e.g. get a 100GB package from Strato HiDrive for 5EUR a month.
(edit:) or 500GB for 10EUR/month.
OK, well let's compare the price without the online backup experience - I can get a 1TB hard disk for £41, so 10EUR a month for 500GB doesn't look like good value to me.
And a 10EUR keyboard has more buttons for even less money.
But seriously: the price for Dropbox consists (invisibly) of storage costs and rent for the comfortable client software. Depending on what value one assigns to that client software, Dropbox may be subjectively overpriced.
Absolutely - hence my snarky response to you comparing apples with apple pie :)
And of course pricing and assigned values are subjective, when but the original article says Dropbox is "overpriced in many ways" even though it's simple, and has saved his "asked" many times, and is now the de facto data sharing platform for iOS, I very much question whether Dropbox is, in fact, "overpriced in many ways".
