What do you think about doing JS-Free sites?

Yes, SIR! Please elaborate what reasonable means in the context of your specific project(s). If that means the software must exactly fit your needs & work flawlessly out-of-the-box, please see below (RFP).
If neither of these fits your requirements, file in a RFP (request for provision) at the <plone-zope-volunteers@pypi.org> developers mailing list ASAP.
You misunderstand me. I'm not looking for help searching the Web, I flatter myself that I'm pretty good at that. What I'm looking for is recommendations for open source software that does the following:

- Calendar/Contacts/Phone file sync (Nextcloud)
- Webmail (Roundcube)
- Wiki (Dokuwiki)

I migrated from Owncloud to Nextcloud after development of the former mostly came to a halt after the fork. Most of your links are for libraries that either depend on an (Next/Own)cloud install or are meant to help you develop a mail front end. I'm looking for a completed, tested product with a medium-to-large community around it. I'll save you some time. I've tried Moin, Redmine, and Trac (and Owncloud as I stated above).
 
You misunderstand me. I'm not looking for help searching the Web, I flatter myself that I'm pretty good at that. What I'm looking for is recommendations for open source software that does the following:

- Calendar/Contacts/Phone file sync (Nextcloud)
- Webmail (Roundcube)
- Wiki (Dokuwiki)

I migrated from Owncloud to Nextcloud after development of the former mostly came to a halt after the fork. Most of your links are for libraries that either depend on an (Next/Own)cloud install or are meant to help you develop a mail front end. I'm looking for a completed, tested product with a medium-to-large community around it. I'll save you some time. I've tried Moin, Redmine, and Trac (and Owncloud as I stated above).
OK I thought you're looking for Plone/Zope modules that do that. I can not give recommendations on these topics. If I understand it correctly, you can integrate PHP-based software in a Plone or Zope web application. I did not say all PHP-based apps are bad. It's just that IMHO Plone/Zope is a better foundation for the reasons I gave above. So when comparing two alternatives, the choice of framework it uses & language it's written in, would be one criterion among others.
 
- Calendar/Contacts/Phone file sync (Nextcloud)
- Webmail (Roundcube)
- Wiki (Dokuwiki)

I don't have a good solution for these but I can share my experiences in trying to solve them ;)

Dokuwiki isn't actually that bad. Perhaps look into the older versions which don't have weird auto updating or quite so much javascript "glitz" to navigate. However you might need to read up on any security issues, luckily almost all of them are related to endless plugins or allowing public write access. If you disable logging in and just ssh / vi the data folder directly, that works surprisingly well (the joys of a plain text / flat file database! take advantage of it! ;))

For Webmail, I personally ssh in and use Mutt. on the mbox directly. If you are expecting a tonne of HTML emails or calendar invites, etc then yeah that is probably a no go. 99.9% of HTML emails are spam anyway but for those sweet individuals using hotmail, then just pipe their stuff through lynx.

As for calendar, etc. I literally have a git repo called "life" that I dump stuff in. Mutt then fetches the alias data from there for contacts.

Whilst I think my workflow is a bit awkward in places, I still much prefer it to faffing with web browsers. Accessing everything through ssh is fairly liberating and security is pretty much guaranteed.
 
OK I thought you're looking for Plone/Zope modules that do that.
I also try to avoid PHP because I unfortunately have some experience hacking in it. However, I've failed to find alternatives to the three pieces I mention.

As I've become aware of my mortality, I've started worrying about handing off my messy personal/family stack to one of my children. Because of this, it simply cannot be composed of a bunch of lovingly handcrafted snowflakes that scratch all my personal programming itches. A reasonably competent person has to be able to Google problems and find answers.
 
I think you can write interesting webcode in ruby,python,php.

I know that lot of webdev(not only basic frontend) for Polish Gov is made in PHP so for sure it is efficient enough. Ruby (on rails??) is used also. I think there is no sense to use Python for frontend.

But all of them will be translated to JS/CSS/HTML before placing on frontend, this is worth to know.
 
What about a language which compiles to javascript ?

This is always the case with professional JavaScript development.

In fact, we should no longer talk about "JavaScript" (but everybody does, myself included), its official name being ECMASCRIPT.
The ECMASCRIPT work group releases a new version of the standard every year since 2015.
Developers usually want (or have) to work with a recent version of the standard (e.g. TypeScript) but today's web browser support ES6 (ECMASCRIPT 6, the 2015 release) and older browsers support ES5.

This is why JavaScript developers use a transpiler, a program compiling a source file to another language instead of object code ; in this case, the target language is an earlier release of the ECMASCRIPT standard (ES5 or ES6 depending on target browsers).
 
Developers usually want (or have) to work with a recent version of the standard (e.g. TypeScript) but today's web browser support ES6 (ECMASCRIPT 6, the 2015 release) and older browsers support ES5.
I personally prefer TypeScript over JavaScript. There's no problem to set target to different version ES3/5/6 ~ E2016-ES2020. e.g.
tsconfig.json
Code:
{
  "compilerOptions":
  {
    "target": "es6",
    "module": "commonjs"
  }
}
 
There's no problem to set target to different version ES3/5/6 ~ E2016-ES2020. e.g.

No problem at transpilation-time, of course. :)
At run-time, your mileage may vary if you target some old browsers (e.g. IE 11, Edge < 15).
For this reason, some companies allow the deployment of 2 browsers on their machines (e.g. IE 11 for legacy applications and a recent version of Chrome for new ones).
However, in a few years, things will become much simpler as the choice of web browser is now in practice limited to Chrome (in the enterprise context).
 
  • Like
Reactions: a6h
For this reason, some companies allow the deployment of 2 browsers on their machines (e.g. IE 11 for legacy applications and a recent version of Chrome for new ones).
For companies with dedicated departments, quite right! Unfortunately I don't have any reliable statistics and I have to express my anecdotal experience limited to me and a few others around me. Pretty much every single web developers I'm aware of (not many!), develop their project with Chrome/Chromium in mind, and a final test on 2-3 platform,namely Chromium(Desktop/Android) and rarely Safari.
 

By itself, WebAssembly cannot currently directly access the DOM; it can only call JavaScript, passing in integer and floating point primitive data types. Thus, to access any Web API, WebAssembly needs to call out to JavaScript, which then makes the Web API call. Emscripten therefore creates the HTML and JavaScript glue code needed to achieve this.

I use JavaScript in my web applications almost only for dynamically manipulating the DOM according to responses to AJAX requests. All the heavy lifting, e.g. site wide searching, math simulations of differential equations depending on user entries, signal measurements, FFT, and generation of SVG plots etc., is done by C subroutines on the server. Now, I conclude, that WebAssembly is not only inefficient and clumsy, it is a non-solution of a non-problem. It is absolutely useless.
 
Lots of interesting opinions stated as facts! ?

Whatever works for you, whatever floats your boat, and whatever gives you the best balance in terms of development time, ease of maintence, security, performance and hits a budget/time frame. Same as programming just about anything, really? :-/

I think the Chrome monoculture is the bigger concern. It's impossible (?) for a small team of developers to build a web browser (in the broad sense of matching what people expect from a "web browser" these days) any more.
 
Joking aside, web development is similar to pop music. It's not some sort of baroque art or underground black metal. You have to listen to the market and client. If you don't you're going broke.
 
This. Who remembers IE6?

I'd say you were referring to the Dark Ages, but that was IE 5/Win and its broken CSS box model (the Tasman engine that drove IE 5.2 on OS X was better). At least IE6 mostly behaved when you used an HTML 4.01 Strict doctype, and there were JS scripts that could be used with IE's "conditional comments" feature to apply some fixes that make it behave a bit better.

What I remember most fondly, however, is the trail Firefox blazed before anybody had ever heard of Google Chrome. Firefox's SVG capabilities, basic XLink support, application/xhtml+xml MIME type support for XHTML, and XSLT 1.0 support were all amazing at the time.

Back then, there was also a genuine desire by much of the Web community to create sites that worked in as many browsers as practical, which usually meant IE6, Opera, Safari, and Gecko browsers like Firefox, Mozilla Suite, and Camino. But as with all good things, it came to an end. HTML5 brought back the "this site may not work in your browser" concept, and Google seems to be the one driving new ideas, so what we're left with are messages similar to those of the 1990s:

This page is best viewed in Netscape Navigator 4.0 Google Chrome.
 
If we depart from HTTP, HTML and its improvements with CSS and javascript, then a new
abstract concept for the web should be invented and consequently implemented.

Perhaps a browser as a virtual machine at which programs from the internet
in a "standard" language for writing GUIs runs. Or should the browser run also
compiler for programs sent? And perhaps the compiler could also be downloaded
as part of the web site?

I think all gets unnecessarily complicated because the development is dictated by
non technicians.
 
  • Like
Reactions: a6h
HTML5 is perfectly fine. It even helped get rid of (or almost entirely reduced dependence of) Flash. Perhaps there's an argument that it needs more features, but it's a markup language, not a programming language. There are so many features a markup language can have, and it isn't intended to be a programming language. It has a clear separation as a markup language, and it needs to stay that way. If it didn't, people would use the same complaints and worse about PHP or Javascript against HTML5.

CSS3 is ok, but it really needs to be improved upon with a newer version or replaced with something else entirely. It takes a lot of foresight to get spacing right, and then each element has to be spaced right again. Large parts of it often has to be redone later, as the website design improves. CSS files are big because each element has to contain specific attributes. Those who've used CSS understand how complex it is for the simplest attributes. Maybe something like Macros would help for specific things like font, and colors.

There is XML based XSLT for stylesheets. I've used them sparingly before, but I don't know how good of a design it can offer compared to what CSS does.

There was another one for stylesheets, something like XSL-FO (Formatting Objects), which I didn't understand because it said that people weren't supposed to edit the file, it was something like automatic. I didn't understand the concept.

I don't understand how anyone keeps preferring Javascript over PHP. Maybe PHP should be replaced, but definitely not with Javascript. Perhaps a good replacement is not invented yet. Or maybe can Python or Ruby replace it? Go-lang for those who like Google? PHP is more universally supported than potentially better alternates by hosting providers.

My .xsession-errors file is full of Javascript errors.
 
I noticed this too. If i quit Xorg, the console is full of Javascript errors. Is there a way to stop Firefox from logging into console?
I don't know. But it's not a big deal. My dmesg is full of GTK related errors anyway. Whenever I shutdown there are a bunch of them showed up.
 
I don't see Javascript errors in my dmesg output, neither with | grep -i java. Perhaps, I have to close Xorg or shutdown to see it there.

They're in .xsession-errors in mine. I could try a shell script to automatically move those specific errors to a new file like .xsession-errors-js.

The file /usr/local/etc/X11/xdm/Xsession sets the errors to ~/.xsession-errors. So maybe I can have a custom script inside ~/.xsession to separate some types of errors.
 
Back
Top