59 comments

[ 3.1 ms ] story [ 97.5 ms ] thread
Just work on making Ghost better if you have a problem with it, why would we need yet another version of it? Because it's written in PHP?
Ghost is really beautiful. I created Ospari just because it is written in PHP and therefore it is easy to install.
Is going to provide all Ghost features or just will look like it?
It is too early to answer this question. The first version would have all basic features needed for blogging. Future features would be based feedback provided by our users.
I would imagine that being written in PHP is the main selling point for end users who don't have the knowledge to set up a Node.js based website.
Yes it is. You don't have to setup PHP. It comes preinstalled and preconfigured on most hosting products. You can copy Ospari and start to blog. No step and no configuration needed.
Also taking in the global context, most hosting plans that are payed in the specific country currency run on PHP without any configuration, just put your PHP files there and puff...magic, a website is running
From PHP to Node and now back to PHP. The circle of code.
No it is not a circle of product. Ghost and Ospari is by no wary comparable to Wordpress.

    >Ghost and Ospari is by no wary comparable to Wordpress.
Regarding Ghost, it depends who you ask[1]. What the grand parent comment is probably referring to is the same thing Rob Conery refers to in that Ghost (in his estimation) was architected the way you would architect a large, modular PHP application.

[1]http://www.wekeroad.com/2013/11/19/hello-minty/

No, but if they work at it, some day they might be.
Not sure how I feel about this. I think we need to move away from the idea of "let's write this in PHP because it's easy." Best of luck to you though...
I think the sentiment is more "let's write this in PHP because it runs on every $7 shared web host on the planet". Sadly, no other language can claim this except maybe Perl.
Well then it's catch 22 isn't it? Unless we boycott PHP it will never go away and hosting providers won't have any incentive to change.
I think you'd have better luck in making ways of supporting drag-and-drop deployment for your preferred languages. I can't find it now but there was such a tool for Python that allowed you to run Python scripts in the same way as PHP (no app server as such).
I don't understand why that is still something people care about in the days of $5 VPS hosting.
I think it's because there are lots of PHP programmers who don't have the first clue about setting up an actual server on the internet (and keeping it up).

On top of that there is the time investment in learning all that stuff and doing it on a regular basis.

To some that makes "here are your FTP credentials, have at it".

----

It's simply one end of a spectrum with DevOps towards the other end.

I agree. I do a lot of wordpress work, and many of the people doing that have no idea what the command line looks like. I don't fault them for it-- that's just what their jobs/business have so far demanded of them.

And while I have setup VPSs for my own projects and they seem to do okay, I've never had to be responsible for a machine that serves other people, and I don't know if I want to be the only guy who has access to a machine that other people rely on to do their business: there are a world of unknown unknowns there, and if I can offload that liability its probably the wiser move.

Bad generalization. There's plenty of developers in general that know nothing of devops and server setup and maintenance. I put it on par with frontend developers who don't know how to slice up a PSD.
>I think it's because there are lots of PHP programmers who don't have the first clue about setting up an actual server on the internet (and keeping it up).

Well PHP programmers are going to want to use PHP anyways. But as far as users not knowing how to setup a server, that's true. But it is also irrelevant. Tons of people provide tons of different VM images for all sorts of things, node included. At that point it is no harder than PHP, with the added benefit of not needing to upgrade PHP every 2 weeks for the latest round of security holes.

I would say more like $2
You are the only one who thinks writing stuff in PHP is 'magically' easy. Let's see your complex PHP projects.
I guess the problem is that many cheap hosting plans (such as Dreamhost) don't run Node but they do run PHP.
This is a huge issue. A good friend of mine runs a webshop that builds custom CMS', marketing websites, shopping portals, and occasionally pretty sophisticated apps for clients. I've talked to him a lot about his industry and PHP dominates because of hosting.

Not only are all the shared hosts really affordable and set up to support PHP, but many larger clients have already invested in their "own" hosting setups, either co-located or with local firms.

He's said there are agencies in his industry that have switched to RoR or Django but that from his perspective a part of their model is the built-in acceptance that the tech will cause friction because of hosting, offset by the offer to host it for a fee...

I'm a WordPress user and haven't tried Ghost...but it seems the impetus of the OP is to port over the "hundreds of beautifully designed themes for Ghost"..I know themes consist of front-end logic and backend code...but isn't the "beauty" of these themes portable independent of scripting code?
This is a solution looking for a problem. Same with Ghost, though. At least Ghost is ironic: a super simple blogging platform that requires fairly sophisticated skills to setup and manage.
The problem is already there. Install Wordpress, buy a beautiful theme and try to configure it. You would be frustrated after 5 minutes…
It states it is open source but I struggle to find the source. Can someone point me to the source?
Yes it is open source and it is free for personal and for commercial use, but we have not released it yet.
Ok, but can one look at the quality of the code before decides to support the project? Is there a repo hosted somewhere?
We have repo, but it is not public

Here a code example.

https://gist.github.com/anonymous/8805880

This is a really weird approach to asking for support for an OSS project. Bad vibes with this one.
Maybe you are right. We would move at least some of our code to Github to a public repo soon.
> Currently we are working on the admin interface

What? Why is there no mention of the live markdown preview. That's what ghost is.

Don't tell me how you've written a database layer, and I can have access to the stupid ghost themes marketplace. I really don't care remotely about that; wordpress has themes. And it talks to a database.

Tell me how you've got an amazing markdown php library with a websocket connection that streams live updates, or even better, how you've made that into a component you can drop into any php website.

-__-

That is really a good point. I haven't written any markdown parser in PHP. There are several Markdown parsers written in JavaScript. I use one of them. Ospari don’t need any web socket connection. The posts are parsed and auto saved as you type and you have a live preview like Ghost.
There are markdown parsers available for PHP (http://michelf.ca/projects/php-markdown/ as an example), but the websocket thing might be quite challenging. PHP's architecture doesn't lend itself well for many concurrent long-running connections.

Unless you want to run out of RAM very quickly, you are much better off to do everything involving long-running connections using node.js or any other environment that works in an event based fashion as threads (and in case of most safe PHP architectures, usually processes) are way too expensive for that.

You are right about the PHP architecture. As wrote there is no need for web socket. You can parse the markdown code in browser.
This is going to end well.
I don't have enough karma to downvote you but I will say that I think your comment isn't adding to the conversation. He's someone with a project that he feels strongly about; misguided or not thought through? Sure.

Instead of sarcasm, help this guy realize his talent for building software could be directed in a more productive fashion than playing "copy cat".

To be fair I was going to explain why it was a bad idea but I find reasoning with PHP developers very hard. As a recovering PHP developer myself, it will just be ignored.
You'll find that with more than just PHP :( I'm seeing it with Ruby and JS / Node too. Some developers are reasonable and great people but there's a subsection that think {Ruby, Python, PHP, JS} are the bees knees. I use all of those languages...
Wouldn't it be easier to port some of Ghost's themes over to WordPress instead of building a whole new blogging platform???

You could even raise funds for that! But I just don't see Ospari going anywhere - the concept is not novel in anyway and there's also a PHP based blogging platform that already dominates a lot of the internet.

Also, Ghost is open source no? I can't find any source files for Ospari...

Node.js + javascript is doing to PHP what PHP did to Pearl in some ways. I see more people pushing that in my circles now than PHP - granted there's a lot of selection bias happening.

Why not build this in Go? I'm sure plenty of people would get excited about that.

[EDIT] If the "key feature" is live previews of markdown, why not fork WordPress and build that feature into it? I'm sure a lot of people would love that - or maybe you can do it as a plugin (I don't know WordPress well enough)??

I just feel like there's a lot of "Not Invented Here" happening with your project, which can be typical of engineering minds. Please focus your talents on something truly novel!

The key issue is the simplicity and the ease of use. Wordpress is not about blogging any more. It is complicated and has a terrible code base. Writing this in GO doesn’t make any sense, because GO is not simpler than Nodejs in terms of setup and installation.
No, I'm pretty sure WordPress is still about blogging - it's got a lot of other features too. Whenever I've talked to a single "average" user they have never talked about "just the blog". Every single person wants to create pages, control the navigation, mess with their templates, install that favorite SEO plugin of theirs, integrate with MailChimp, have a login system, sell ecommerce products.

Even Ghost is acceding to those needs with their Pages features and I'm sure soon to exist plugins (if they don't have it already).

WordPress' codebase is not terrible. Kinda messy? Sure. Idiomatic? Not so much. Does it work? Yes. Does it power millions of websites and blogs? Yes.

My point with Go is that if you want to have a successful project you have to do something novel and there's nothing novel in what you're doing - you're an engineer experiencing "I can do this better" syndrome. I do that all the time, take a step back and really think hard about what my value proposition is (even for OSS projects). If it's already being served by some other project or product and I want that thing to be "better" then it makes more sense to help improve the other thing.

Or in Ghost's case, build it on an entirely different platform (which is novelty), with novel features.

I don't see how running a ghost instance isn't easy or cheap, there are a lot of options: http://www.howtoinstallghost.com/
Interesting project but I'm a bit turned off by how this is all framed.

"We're building a new blogging platform that can use Ghost themes, check it out" comes across a lot more positive than "We're porting Ghost to PHP, please give us money to do this."

Thank you. I was looking for catchy headline. This headline didn't come to my mind. On our actual campaign we don't use this title.

We would do it, even if no one gives us money. Some money would just accelerate the process.

Update: We just cleared the licensing issue. Ospari would be released under MIT License.