4 ms·

An issue with projects like AMP is you are now essentially solving a completely self-inflicted problem. Nobody required the particular approach that AMP takes,

An issue with projects like AMP is you are now essentially solving a completely self-inflicted problem. Nobody required the particular approach that AMP takes, and now site owners have to worry about blank pages for the first time since 300 baud modems? The solution needs to start in much simpler terms: Google should publish page traits that are rewarded and punished, and tools for seeing where you stand. It should reward/punish all sites based on how they fare in the published areas. We don’t need new protocols/wrappers/rehosting or extra scripting or whatever else to drag bad sites into the 21st century. These just create additional issues when we have plenty of engineering problems already.

jorblumesea 8y ago
It seems far easier to establish a standardized framework that forces everyone to adhere to it as opposed to trying to get the web to adhere to some nebulous perf standards. The web has incredible variety in tech stacks, business structures and even views on what a standard means.
- danShumway 8y ago
  > The web has incredible variety in tech stacks, business structures and even views on what a standard means. But AMP is a standard. Using AMP doesn't solve any of the problems you're talking about. If people's usage is too diverse to be accurately measured, then it's too diverse for AMP to meet everyone's needs. And if it's not important for AMP to meet everyone's needs, then why is it important for a set of page tests to do so? I agree there might be problems building a small set of tests to check page speed, but that would still be strictly better than AMP. It would still cover at least all of the use-cases that AMP covers now, and it would open the door to cover more use-cases in the future. If you're ranking search results based on whether or not someone uses a framework, you are implicitly ranking them based on the attributes of that framework. What people are asking is for Google to make those attributes explicit instead, and to directly test them.
  - growtofill 8y ago
    > But AMP is a standard. AMP's website defines it as a library: https://www.ampproject.org/learn/overview/ https://www.ampproject.org/learn/overview/
    - danShumway 8y ago
      Meh. For all practical purposes, it serves the exact same role as a standard. It's a set of rules/technologies that developers have to follow in order to get preferential treatment in Google search. I don't particularly care whether anyone thinks that's technically a standard or a framework or a library by the ridged definition or not; at the point we start down that rabbit hole we're just talking about words, not concepts. What I was trying to get at above was that any problems people bring up around Google profiling websites for search placement are still present in AMP. Forcing developers to use a specific set of technologies is functionally the same as forcing them to conform to a ridged set of benchmarks. For the purposes of this discussion, we might as well call AMP a standard.
  - jorblumesea 8y ago
    Precisely, it's a ways of enforcing a standard via a framework. Far easier than setting a standard and trying to make everyone follow it.

jorblumesea 8y ago

It seems far easier to establish a standardized framework that forces everyone to adhere to it as opposed to trying to get the web to adhere to some nebulous perf standards. The web has incredible variety in tech stacks, business structures and even views on what a standard means.

danShumway 8y ago
> The web has incredible variety in tech stacks, business structures and even views on what a standard means. But AMP is a standard. Using AMP doesn't solve any of the problems you're talking about. If people's usage is too diverse to be accurately measured, then it's too diverse for AMP to meet everyone's needs. And if it's not important for AMP to meet everyone's needs, then why is it important for a set of page tests to do so? I agree there might be problems building a small set of tests to check page speed, but that would still be strictly better than AMP. It would still cover at least all of the use-cases that AMP covers now, and it would open the door to cover more use-cases in the future. If you're ranking search results based on whether or not someone uses a framework, you are implicitly ranking them based on the attributes of that framework. What people are asking is for Google to make those attributes explicit instead, and to directly test them.
- growtofill 8y ago
  > But AMP is a standard. AMP's website defines it as a library: https://www.ampproject.org/learn/overview/ https://www.ampproject.org/learn/overview/
  - danShumway 8y ago
    Meh. For all practical purposes, it serves the exact same role as a standard. It's a set of rules/technologies that developers have to follow in order to get preferential treatment in Google search. I don't particularly care whether anyone thinks that's technically a standard or a framework or a library by the ridged definition or not; at the point we start down that rabbit hole we're just talking about words, not concepts. What I was trying to get at above was that any problems people bring up around Google profiling websites for search placement are still present in AMP. Forcing developers to use a specific set of technologies is functionally the same as forcing them to conform to a ridged set of benchmarks. For the purposes of this discussion, we might as well call AMP a standard.
- jorblumesea 8y ago
  Precisely, it's a ways of enforcing a standard via a framework. Far easier than setting a standard and trying to make everyone follow it.