• Activity
  • Votes
  • Comments
  • New
  • All activity
  • Showing only topics in ~tech with the tag "web". Back to normal view / Search all groups
    1. Copyleft software license is broken when it comes to web apps

      Now, we all much admire the Richard Stallman mindset and the libre/commons ideals of GNU GPL folks, they are the ones who pioneered the open source foundations by giving us things like core utils,...

      Now, we all much admire the Richard Stallman mindset and the libre/commons ideals of GNU GPL folks, they are the ones who pioneered the open source foundations by giving us things like core utils, emacs, gcc, etc. at a time when proprietary solutions were having a field day with absolutely zero competition.

      However, the GNU GPL software license has a few practical problems when it comes to development of web apps. I came across these while developing a PHP web framework recently. I had initially considered GPL v3 but I will have to go with a permissive license like MIT or Apache due to these issues:

      1. The GPL applies to your entire software as a whole including derivative works. Now what constitutes a derivative work is often highly technical but not even an attempt has been made in GPL to clarify that. That question has been left vague for some reason which is never good from a legal perspective.
      2. In my case, the two require files are in the core directory while a plain index.php lives in the root for the user to override. This index.php is included as a template or stencil which the framework's user is supposed to override with their own code. In GPL lingo, this might well cause it to be a derivative work and that will require them to "contribute" those changes back to me which doesn't make any sense at all! There needs to be an exception for included sample or example files for which the GPL shouldn't apply.
      3. There is also a problem with GPL due to the nature of web or Internet. Another app I am developing is a bit heavy on JavaScript and contains static *.js and *.css files. The mere running of the app in a web browser will cause these files to be "distributed" through the <link> and <script> tags. From GPL perspective, this becomes a technical violation as no GPL license had accompanied this distribution. And in case you minify or compress your JavaScript or stylesheets for efficiency (which is a very common practice), this violation becomes even more grave as you're technically distributing GPLed source in an obfuscated form!

      The GPL badly needs to be upgraded for these scenarios. Until then, it remains a good use case for desktop or console apps which run directly on your computer and usually have a clearer boundary of what constitutes a distribution and/or derived work.

      14 votes
    2. Unpopular opinion: Disqus is the Internet's welcoming gift to the small and medium bloggers

      I'm talking about those who host their blogs statically on Github Pages and elsewhere, the small peasants of the Interwebs who collectively make up the biggest portion of it and make its wheels...

      I'm talking about those who host their blogs statically on Github Pages and elsewhere, the small peasants of the Interwebs who collectively make up the biggest portion of it and make its wheels churn with their content and toil.

      All that negativity I'm seeing towards Disqus is baseless. They criticize it for showing ads and being less privacy friendly but then which technology today isn't? Name me one company from Big Tech or Silicon Valley which actually respects their user's privacy in practice, not just words. You'll be left scratching your heads for a long time but no name will come out.

      And I also keep hearing about these "self hosted" alternatives like Commento but none of them have gained any traction. Building and implementing a commenting system isn't easy even though CMS like Wordpress try their best. Unless you implement social OAuth logins like Github or Google, not many folks will like to interact on your site by singing up for a whole new site account, especially true if you're a small blogger who has just started. And managing all the data communication between the frontend and backend isn't easy.

      With Disqus, we get this whole setup and workflow readymade with nice looking dashboard and all, and that too for gratis, not a dime to pay! Then why shouldn't a blogger use it, I'm yet to hear any justifiable answer to this question.

      8 votes
    3. For those involved / interested in Web3, what do you make of the near and long term future for it?

      Added the qualifier to the title as web3 understandably earns a lot of eyerolls haha. At the same time, a lot of web3 focused places seem to have a specific mindset about what "should" be done so...

      Added the qualifier to the title as web3 understandably earns a lot of eyerolls haha. At the same time, a lot of web3 focused places seem to have a specific mindset about what "should" be done so I wanted to ask here.

      I worked in the space at startup (ironically making web2 services to assist in web3 so I’m still an extreme novice). But my time there was a constant push / pull between convention and money and innovation and the unknown. The company I was at would try to appeal to big companies in hopes of finding a product market fit, who looked to us for guidance on what to do in this new space where they hoped to make money. Trend after trend would pass and it would be entertained whether we’d jump on it because product market fit.

      The most desirable companies were household names with non-web3 userbases because they meant unprecedented reach. But to make web3 approachable to them, you’d have to define a UX that didn’t exist and would be pulled in a tug of war between two forces. The first mindset optimises for the purest idea of giving the user power— UXs that were obvious about the concepts of transactions and transferrable assets. The other wanted to replicate web2 UXs in web3, to the degree that a user gives temporary control of their wallet to a developer so the developer performs transactions as them.

      Then, there is the data and pseudonymity piece. Companies have been taught that data is valuable, and one of the values of a blockchain is an identity that exists outside of any one company. But if all of your assets are on a blockchain— either under your public key or perhaps under a few that might transfer assets only between each other— then your identity can be known (not so private) and also cannot be monopolized and sold (because your data is public).

      In the background, as this all happens, is the decentralization argument. At the end of the day, my company used EVM nodes operated by another company (which themselves might be wrappers around something offered by AWS). What is meaningful decentralization alongside specialization of labor? What is decentralization in a world that has billionaires and enormous companies who has the means to buy resources and set up tons of nodes?

      Being out of the space now, I do think a decentralized database with immutable scripts, user-managed transferrable assets, and transferrable identity has enormous value. But recently I’ve been wondering how much of that can be accomplished in the private sector. In my time there it felt like the startup needs (enterprise customers, increased ARR) constantly compromised the will for innovation efforts.

      19 votes