Hacker News new | past | comments | ask | show | jobs | submit login
Publii: Open-source local WYSIWYG static site CMS (github.com/getpublii)
268 points by api on July 23, 2022 | hide | past | favorite | 75 comments



It's interesting to see people rediscover this idea of "develop locally, push static site to web," but I think the weirdest thing for me is, why was this not a continuous thing?

Broadly, this looks good for a blog format, but not sure if that's what people need?

Anyway, for those interested, I just use http://zim-wiki.org plus a custom CSS template I did.


> It's interesting to see people rediscover this idea of "develop locally, push static site to web,"

Yes! Dreamweaver was good 19 years ago and remains good today. Being able to see what you're doing, without running like treacle in the browser whilst editing (I'm looking at you, Divi and other WP visual theme builders), is an awesome experience. Like professionals in many other fields expect from their software.


I used dreamweaver when they had coldfusion integration. We ended up using the wysiwig just for verification because it would make the code so bloated (by late 90's standards).

I recall using HotDog Professional for my personal stuff. Anyone else use it?


I remember using HotDog pro back in the day. I don't remember much about it though, even how much I used it.


Yup, I used HotDog Pro for a while and moved onto HomeSite (https://en.wikipedia.org/wiki/Macromedia_HomeSite) at some point.


I remember being one of the first users of Hotdog. Sending heaps of bug reports and feature requests, and in the end even having the author over in Belgium to negotiate a European distribution deal.


Many other fields do not have to edit turing complete results. Also you don’t need WYSIWYG but WYSIWTG, since you want to see what THEY will be getting, they being endless combinations of systems.

I was there, too. Dreamweaver was not good enough.


Dreamweaver all too often spun nightmares


> Dreamweaver was good 19 years ago and remains good today.

Notepad.exe was good 19 years ago.

Dreamweaver? Not so much.

:)


Dreamweaver was a good idea, but impossible for them to maintain. In my experience it was regularly giving bad results.

It was a great tool for learning HTML and CSS though, but mainly because you would always end up just going into the source to fix it and view in DW.

I would be very nervous of anyone using it these days for anything else.


I'm actually curious to see what Dreamweaver looks like these days, because I just recently found out that it's still around. I haven't seen it since like 2009 and I honestly thought it had been shut down.


It's still around, but it's no longer being improved, just getting maintenance. Most of the people involved with it migrated elsewhere.

Most who used Dreamweaver and relied on its extensions are now using another similar app called Wappler.


For those that are into Bootstrap then Bootstrap Studio is also a good option which I found produces pretty decent results.


Yep, I'm a fan of Bootstrap Studio, if for nothing else than the fact that it's not a subscription product.


>Dreamweaver was good 19 years ago

Good for what ? I remember designers trying to create pages with and the generated code was useless garbage once the customisation requests came in and they called in a dev


I would say Dreamweaver was OK, but it was eclipsed by the web-based CMS. Dreamweaver generated trash code when it was used as a DHTML design tool, but it was never as bad as FrontPage. When used in design-mode it would build up a model internally, and then try to generate code that it could later roundtrip. This worked when things were isolated, but as complexity grew it stopped working and was hard to debug.

I saw it work reasonably well when someone took a code-first approach with the overall website templates which avoided the mess from building these up using DW's code generation. You could then set it up so people could only mess with the certain parts of the page which avoided the overall site getting messed up as you added others. This got cleaner when they introduced Contribute.

Dreamweaver was created for a time when there was still a webmaster-type role in companies. These people didn't necessarily have what we'd consider a strong dev skillset, but the world around them had started to change which explains the range of capabilities in Dreamweaver.

If you want to see some terrible server-side code generation, you should checkout Drumbeat (https://macromedia.fandom.com/wiki/Macromedia_Drumbeat).


>why was this not a continuous thing?

Because we wanted Dynamic webpage back then. Or at least dynamically generated webpage when something is updated, like links and comments ( When every site was still hosting their own comment sections ). Develop Locally and push Static Site to web may work on Web pages, but doesn't really work for website with some traffic or what we called blog today.

Especially when we were on HDD and single core CPU, updating a few hundred pages takes quite a bit of time. Before uploading tens of Megabytes of Data that takes hours on 56K or ISDN. Compared to being hosted on the web where everything was a press of a button.

Now we are on SSD, where our IOPS is anywhere from 100x to 1000x faster. Even updating sites like Tomshardware or Anandtech should only take seconds. We have Gigabit Internet, uploading should no longer be a problem.

I think vast majority of web site could simply be Static pages, with a small dose of javascript. And may be in the future you could even install your CMS on iPad or iPhone. There is no need to worry about anther Tumblr or Geocities taking down all your content. Your content will stay on your device, and possibly with some sort of Cloud Data Backup. ( iCloud or Google Drive )

This is actually something I keep talking about since ~2017. ( I cant believe that is 5 years already ) So I am glad we are finally moving in that direction.


> why was this not a continuous thing?

We had the likes of Netscape Composer, Front Page and Dreamweaver. They produced mountains of un-maintainable and not non-compliant html code.

When HTML5 came along there was a big push for compliant sites. Also it made sense to have a developer use the power of CSS3 to handle layouts and responsiveness elegantly. So the WYSIWYG editors fell out of favour.


As opposed to what we have now? :)


I always preferred it that way myself. Used netobjects Fusion for years and many sites.

From what I remember, the push for html5 and super lean / clean / and 'validating code' for google search results made it difficult for the older tools to stay in favor.

Also the move to mobile.. and not just "m.site" - making so javacript / bootstrap type things came to be favored for a while (lack of grid / flex in browsers at the time) - which mostly required text editors instead of wysiwyg.

the business models for investing in software were shaping cloud things around that time (some say for better, I say for worse) - so webflow, wordpress, adobe cloud everything, and similar had a chance to do the things that stand alone download and own software could of been doing.

Faster internet being more prevalent and such of course helped with these things.

I've been itching to spend a couple hours testing pinegrow site building ( https://pinegrow.com ) to see if it is reminiscent of the netobjects days.. and now that wordpress is getting increasingly painful to use and maintain, I'll likely get to that sooner than later.


I think that user experience of website builders with wysiwyg and drag and drop UX won over time. Then, as time passed by, website builders become bloated and complex. Once again you needed a professional to maintain your site in site builder.

So now simple solutions, static HTML, free or one-time fee CMSs are sexy again. (economy is not good, who wants another subscription?)

I know because 14 years ago we have created static, drag and drop CMS that people still buy today. People value basic features. And it's hard to make a complex software that's easy to use.

Basic features are still enough, after 14 years. https://sitecake.com


This might be good for a personal blog where you have it local on your machine, but want a blogging interface that's separate from your coding tool (VSCode, etc).

The problem is that if you have to pull down the site to edit content, it defeats like 99% of the use cases of a CMS. The whole point of a CMS is to provide an interface for editing a website without editing code directly and not having the code on locally to make edits.


I mean, it's been pretty continuous? Dreamweaver to Hugo to Next/SvelteKit/etc. is a pretty straight line.

The inability to have interactive data-based content in static sites is why I went away from them, but the ability to progressively enhance a Next or (especially) SvelteKit site with data-y bits when it makes sense has gotten me in the habit of using them. Kind of a fan.


Would be interested in seeing your website if you might care to share it?


http://jrm4.com

With this:

https://github.com/jrm4/Eight-Five-Zero

(Kind of messy, but it's utilitarian for my work and such)


Thanks. A lot of good resources on your site!


Thanks, I think the thing that has served me the best is how easy and quick I've made it to update. I open zim (which is a native app, no login or anything), change what I need to change, and have a short little SSH/rsync shell script that updates.

It's really nice that I can update the website in the middle of class as I'm teaching with very little issue.


Wow, seriously nice work. Downloaded it and started playing around. Pretty much exactly the features I wanted from a CMS. Both simple and powerful.

I attempted to build a CMS years ago and this is pretty much exactly what I wanted to create... but sadly was too incompetent to finish.

Nice initial list of themes, hope they build up a designer community around this.

I could see a commercial hosted version being pretty popular as well (for companies that need to collaborate on posts). Easily could outcompete Squarespace and others.

Anyways, bravo!


Huge Publii fan here, glad to see it on the front page.

I feel like the only thing holding Publii back is the lack of theme diversity, and definitely could use some third-party designer love to push it forward.


This seems to be the story for nearly all static site generators (except for maybe Jekyll, whose drawback is it's a hassle for non-Ruby devs to properly setup a Ruby environment, especially on Windows).

I migrated from WordPress to Hugo for awhile, but eventually gave and migrated back. You can love WordPress or hate it, but it has an infinite variety of paid and free professional-quality themes. Hugo has a slew of one-person amateur projects, the best of which are half-baked clones of WordPress themes. What's worse, they tend to be poorly-maintained, and stop working whenever Hugo introduces breaking changes (which at least a few years ago, was frequent).

I appreciate all of the conceptual advantages of a static site generator versus a CMS application host. But for most users, a website tool is simply only as good as its theme ecosystem.


I agree that this does not yet replace wordpress, however it's a step in the right direction.

I do think that this does have an audience, travelers or those in places with bad network infrastructure like volunteers in the Peace Corps could make good use of this, where they may go months without getting internet access.

Also, I see this as a great tool to use for my personal non monetized blog, where I don't need the infinite variety of themes or the need to constantly keep the system up to date. I just want something I can type up notes or simple ideas and publish them simply. I don't think this is a huge subset of users right now, but it would be nice to have things like this more available for this type of use.


Yes, more themes are needed, but I think they should also focus more on options for creating pages, especially landing pages. This would be very interesting for business-focused sites — which I believe are the majority these days.


This was my take away too. Looks great, but personally if I'm going to use this for a docs site, or for a wiki, or some sort of knowledge base, I'd really need to have an out of the box theme available. There are some there, but not enough variation for my liking. So rather than a quick win, this has to be put on the pile of "take a look and determine if I have time".

Not that it takes away from what looks like a great product, but that's the journey I went through, as a regular, overly-busy employee-startup founder, 14 hours into a Saturday. "I love it, but it's over my time-cost threshold if it needs custom themes".


I use Publii for a blog and enjoy it. I wanted actively developed static site generation with some reasonable template/theme options, and I didn't care about multi-user or team collaboration. I achieved multi-device sync via storing Publii in a cloud location.

I looked into developing a theme using their theme system and it seemed unattainable for anyone who doesn't have significant time to invest in understanding the guts of Publii.


Does Publii still not officially support git versioning? If it did I probably would have played with it..


You can create a regular Publii-independent Github repository to hold locally generated versions of your site.

This repository can then be automatically synced to the live site via Cloudflare Pages. This option even allows using a private repository.

https://getpublii.com/docs/configure-cloudflare-pages-with-p...


I don't think Publii is aimed at people who know what git is. Those people already have lots of great options.


It doesn't have to be git as much as just some form of versioning. Everybody loves versioning, they don't have to care about what powers it.


do you know if tailwind can be used in it?


It can. You can use whatever CSS framework or library to create themes.


There is a somewhat similar system called Lektor, originally written by Armin Ronacher (of Flask/Jinja fame): https://www.getlektor.com/ You define your models, then start the local devserver to add entries for the models. In the end, it stores the data in the filesystem and outputs static HTML.


I scrolled down, hoping to find mention of Lektor. It is an outstanding and under-appreciated system. I have been using it for all my sites for several years.


Good to see the MS FrontPage days are back :]


FrontPage was ahead of its time


This is the first time seeing this and I have to say I'm super interested.

Been researching some CMS' recently and was not really impressed, was thinking of a Frankenstein with Strapi but this looks like it could be absolutely perfect.

Great work, will check it out further and so far from what I see will be joining the community to help !


Great to see Publii featured on HN. A few years ago, I switched from WP to Publii for my personal blog. Here's a post I wrote about the experience: https://www.callmefred.com/i-switched-from-wordpress-to-a-st...


I always liked this model of personal CMS. Back in the day i think it was called MovableType or something was quite popular too.


running several blogs on publii! (https://typogram.co/blog/) Loving it!

the only feedback I have is making co-editing easy. sometimes I co-write articles. this is rly hard with a desktop app.

Also, wishing for some kind of SEO plugin that would make writing descriptions easier.


Did Publii ever get nested structure support? I remember having to perform workarounds for subcategories.


I want a love child of this and Elementor. That would be perfect https://github.com/elementor/elementor


I don't think this will happen in the near future. The app doesn't even support pages properly.


I've been using this for a while, not perfect but pretty good. I'm enjoying not having to deal with security updates for my simple single user CMS. While being able to quickly and easily add posts and articles without messing about too much with web dev administrivia.


I've been looking for something like this / thinking of developing a light-weight, similar tool.


Has anyone experimented with its WP importer and could comment about its quality or efficiency?


I used the WP Import feature when I switched from WP to Publii for my personal blog. It worked fine: https://www.callmefred.com/.

All details in this post: https://www.callmefred.com/i-switched-from-wordpress-to-a-st...


Not very reliable. There are many reports of problems in the public forum: https://forum.getpublii.com


Seems to be working well for my small blog. Only issue is pages don’t import well, I would export only your blog posts.


It works. But it only works for standard WP stuff. Throw in a non-standard gallery or something, and it will degrade.


Wish something would exist as a GUI editor for other static generators, like Hugo.


Try Quiqr, which just launched at HugoConf:

https://open.quiqr.org/

Fair warning, it’s a bit rough around the edges.


> Fair warning, it’s a bit rough around the edges.

Gave it a try - imported my hugo blog but then won't show me any of the content, just says "add content to a new site!" Hopefully this is an easy thing for fixing (I suspect the antiquity of my hugo blog might be confusing the issue.)


is this gui? you can edit in wysiwyg?


Yes, it's a WYSIWYG front-end running locally on your PC, Mac or Linux computer.

Mind you, I haven't actually used it much beyond importing a WP site, so this isn't an endorsement.


totally agree, did a deep dive into this and the keywords kept bringing up webflow essentially.


Reminds me of the product that Fog Creek Software was originally founded for, CityDesk.

And another tool that got some traction back in the day: Radio Userland.


Cool, but it's only for blogs.. I was hoping for something similar with the concept of "pages", and design-blocks


I too was surprised at the lack of pages, however it does seem to support some kind of design block system.


static site generator + visual studio code is good enough. vscode has emmet that can create <html> part. vscode also has image preview extension. I used to create my own GUI, but find out it was unnecessary


my impression of cms means you need login to read the content except for part of them that are set to public,static site is not a cms per that definition to me.

it is perfect for sites contain all public content though.


Looking at the source takes me back to 1999.

Nothing wrong with it. Just a mix and serves a purpose.

Thanks


Looks awesome! But why not providing a web interface as well?


Great as mini-solution, appreciate it.


The return of dreamweaver...


Dreamweaver was a visual code editor, Publii has no resemblance to it.


WYSIWYG's are essential to have non-coders post on your site :/ Or for handing off to mom-and-pop clients.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: