tored

Tab hoarding is not an effective strategy of organization, it is more the fear of losing something important. And tab hoarding builds up, just like any other type of hoarding.

Finding tabs was a bit of hit or miss, you can search for them if remember the title, otherwise you have to manually find it, time consuming.

With my new strategy that I described above, it has become much easier to find what I’m looking for.

Bookmarking has always been clunky, putting them into hierarchical folder structure is both time consuming and has the same problem as files and folders, it lacks a dimension, a bookmark can be in multiple folders.

Bookmark toolbar could work, how paradoxical it may sound with all the tabs opened, it just don’t like that UI element cluttering my browser.

Forgetting about bookmarks can be a good thing, it is similar when browser loses all opened tabs, the feeling is a strange mix of anxiety and relief, anxiety of missing out but a relief that I don’t need to process them anymore.

Sometimes a find ten year old bookmarks on some backup, most bookmarks dead or just out of date. I feel we need better ways to organize things in this temporary cloud world.

My problem with bookmarks is that I tend to forget about them.

What would be nice is a browser where tabs, history and bookmarks blended together seamlessly.

Probably because snaps autoupdates without user approval.

Can I backup my container configuration to a json file on disk or do I still need to sync it to the cloud to save it?

Yes, it is the 200+ tabs (probably closer to 500) over multiple windows.

My latest move was to merge every tab in all windows to one window only (with an extension) and start using vertical tabs (better scrolling and overview of tabs) Then sort tabs by title with another extension (Edge built in AI tab sort sucked on sorting so many tabs). With sorted tabs I could start create tab groups, Edge AI tab sort worked better when the tabs was already sorted alphabetically and managed to create most of the tab groups for me. With all this reorganization it was easier to manage all of them and to start closing tabs.

I’m not done yet but it is much better and number of opened tabs has been significantly reduced. Now I have one main window with all the tabs and some temporary windows that is only used for temporary stuff that get closed within a day.

What makes things a bit more complicated is that I also use two profiles, private and work. Firefox always sucked on multi-profile setup. Firefox’s new container stuff is somewhat improvement but not fully (at least when I tried last)

I find Firefox much more heavy on resources vs Edge. I’m always get disappointed when trying to make Firefox my main browser.

Chromium devtools has more features but more cluttered and more annoying to work with.For the common devtools tasks Firefox works better IMHO. But that can be my bias after using Firefox/Firebug devtools for over 15 years.

I find your answer most interesting yet, especially about Article II.

Apparently the president can appoint temporary advisors that does not need approval from congress. If Mr Musk qualifies for such appointment I guess is that something for the courts to decide. I guess that implies that the audit itself is time framed.

It seems that many objects to the audit on technicalities and use that as an argument that the audit itself is illegal or unconstitutional. It is a flawed argument.

If they have security clearance then the audit would be legal?

As long as they keep the data safe the audit is legal?

US president has a lot of powers, I’m not aware of any elected official in Europe with the same amount of powers (ignore Russia).

President of France is probably the most comparable, but in France you also have the prime minister, selected by the president but supported by the parliament.

In Sweden we have a separation of powers within the executive branch. Government agencies are independent of the cabinet.

DOGE’s audit wouldn’t be possible in Sweden, that would have required legislation or even constitutional changes.

Sweden has already an independent government agency that audits the rest of the government, but it has support in the constitution for that and it is technically administered by the parliament and not the cabinet.

https://en.m.wikipedia.org/wiki/Swedish_National_Audit_Offic...

My personal opinion is that the US system of government is vastly inferior to Sweden’s.

So if DOGE have security clearances (unclear if the have) then their audit is legal?

Dependency injection containers can be overly complex and allow complex configurations, too much magic.

That is why I kept my container library simple, you read thru all the source in less than 15 minutes.

What I like about it is that it is natural to use composition with classes, it encourages to create classes to divide your application into reusable parts.

And with my library all you do is to put the dependent class in the constructor. My library can without much effort be replaced with normal factory, that is by design.

Warning PHP https://github.com/paketphp/bero

Important to note that USA is a republic, typically in Europe parliamentarianism.

Which laws? The article describes security clearance.

Probably not.

https://en.m.wikipedia.org/wiki/Disposition_Matrix

Of course you can. Write your own abstraction of Unicode string, it is easy. Php has all the tools for it.

Yes, in mainstream PHP you use the webserver to leverage forking. The great thing about that is that vastly simplifies the language and any frameworks written in that language. No, it doesn’t force anything to the client.

Also any performance gains made in web server technology is given to PHP for free.

In Java and Python and the like you typically have a thread pool to handle every request, this has the benefit that you can share memory with mutexes between request (threads) . However 99% of cases that is not typically something you want, to have leaky requests, thus most of the time you write code similar to PHP to avoid any hard to debug lock. In Python you also have bad threading performance because of the GIL, thus you avoid anything that steps into that.

Thus PHP share-nothing-model wins again. It is also a natural fit to stateless REST. In PHP you use the database to share memory (data), which is a natural fit in a web stack. And as with the web server you can now leverage any performance gains in database technology.

Nodejs is of course complete trash with the mini freezes of the event loop. And to leverage multiprocess you have to use something like cluster, thus you are back to share-nothing again.

Thus what you argue for is to change a successful request and memory model because of that 1% use case. That is dumb.

How?

Yes, because functions can’t be autoloaded and you can’t put variables and constants inside a namespace. But it isn’t hilarious, it is because of legacy.

Now there are proposals to fix function autoloading but it won’t fix variables and constants.

One way to do it better is to implement modules.

This can definitely be an issue and it was when Guzzle released a new major version every month.

Also when working with projects that are 10+ years and lines of codes a million you will run into this. Sure you can always fork and move the library to another namespace, but that is extra work.

Modules are vastly superior to namespaces, with modules it is the caller that decides the symbol, with namespaces it is the callee.

Another problem with namespaces in PHP is that you can’t put anything into it, like a variable, it is not like namespace in C++. Namespace in PHP is just name added to the class or function name, kind of fake. Don’t get me wrong, it was a pragmatic solution to a real problem and it worked for what it tried to solve, make it easier to share code.

Modules could be elegantly solved in PHP by reusing the internal concept for a class (how it is with enums and interfaces)

Modules would also make it easier to unit test code, because you will not need to worry about a symbol already have been loaded and can’t get unloaded.

With that said I don’t think it the most important feature to focus on, there are others that give more bang for the buck, like string templating and operator overloading.

It is not a non-issue. You can run into naming collisions by just using two different composer dependencies where each dependency depends on the same library but different major versions.

It becomes a problem if you want to use multiple versions of the same library, which may happen if you are using composer, e.g two different dependencies uses the same dependency but different major releases.

How does any of these languages help me write a proper HTML document?

PHP has unicode support which what you need. You only need to understand when to use what function.

Concurrency support is not needed because that is what the webserver is for.

A great bug fix release on a great major release. Thank you to everyone involved to make this possible.

You could almost use the same coding style, typically you don’t what to leak data between handler threads in Java so you can write in similar fashion where you start from scratch at every request.

But Java has more ceremony with the stricter type system.

A bigger problem (at least with older Java versions) is string building and templating. You typically do a lot of string stuff in web development.

Deployments and dependency management are worse with the wars and the jars and how to build them.

PHP is closer to the metal so to speak where many APIs are thin layers around various C libraries, e.g I find working with database (JDBC?) in Java annoying. I feels like Java is it’s own island in that way.

Java’s plus is better performance for CPU heavy tasks, connection pooling and easier when you do things asynchronous. Streams are a plus too. But those things are not important enough to pick Java over PHP IMHO.

Thus Facebook is not censoring Linux discussions or Linux content, what DistroWatch claimed, it blocks linking to what Facebook deems as malicious links (correctly or incorrectly), something a lot of software does these days.

This is what the yanks call "a complete nothing-burger".

Yes, you can read about it the rfc under "Abbreviated syntax"

https://wiki.php.net/rfc/property-hooks

It follows the syntax of arrow functions

https://www.php.net/manual/en/functions.arrow.php

Great PHP release. Better stack traces for closures, performance improvements (always nice), HTML5 support, lazy objects and much more. Great work and a big thanks to everyone involved!

I work on multiple projects from PHP 5.2 to PHP 8.3 and everything in between.

Statistics based on packagist(composer)

https://stitcher.io/blog/php-version-stats-january-2024

Statistics based on web servers in the wild

https://w3techs.com/technologies/details/pl-php

https://w3techs.com/technologies/history_details/pl-php