Work on HTTP 2.0 is offered to start over.
Open source developer Paul-Henning Kamp (Poul-Henning Kamp) appealed to the members of the HTTP Working Group with a call to throw out current developments under the standard HTTP 2.0.
Paul-Henning Camp is the author of MD5crypt and a large number of system components FreeBSD, GBDE, UFS2, malloc, and so on. He believes that the HTTP working group should admit defeat - and start all over again.
As an exemplary fiasco, Paul-Henning Kamp cites the example of SPDY. In classic works on project management, it is said that “the prototype system should always be discarded”, here Kamp refers to Frederick Brooks and the book “Mythical man-month or How software systems are created”. According to him, the SPDY protocol adopted as the basis for the HTTP 2.0 specifications is exactly the prototype.
SPDY was accepted even before the working group completed its previous assignment, and then spent a lot of time to bring SPDY to mind, correcting the shortcomings and errors.
')
And then everyone suddenly discovered that SPDY doesn’t even solve a lot of important problems for which they have to resort to some simplification of the HTTP concept itself.
Kamp's letter was sent in response to the acknowledgment by the chairman of the working group, Mark Nottenham, that "we may not be able to do everything correctly in HTTP 2.0, and we have not yet coped with everything," and therefore decided to "start a discussion on HTTP 3.0." Nottenham emphasizes that the working group is working on tight deadlines, in a hurry.
“Now even the chairman of the working group has publicly acknowledged that the result of the work is a partial fiasco and that we have to replace HTTP 2.0 with something better“ soon, ”says Paul-Henning. - So what exactly do we get from continuing this work? Maybe it would be better to consider the current situation with cryptography and protection of private data much deeper than to publish a protocol with a cryptographic patch that does not solve problems and interferes in many applications? ”
According to the developer, the adoption of the HTTP 2.0 standard just because the deadline has arrived and it needs to be done according to the procedure is not necessary for anyone, it only takes time away from everyone, leading to additional security risks without any significant benefit.
“Wouldn’t it be faster to start solving this problem - creating a protocol that _can_ replace HTTP 1.1 in all scenarios and really will be an improvement in _vseh_ scenarios?”, Paul-Henning Camp asks a rhetorical question. He calls SPDY to be considered an interesting prototype, which clearly showed the need to improve HTTP 1.1, but immediately set about developing a new protocol that replaces HTTP 1.1 (taking into account all the developments of SPDY).
Paul-Henning Camp is the author of MD5crypt and a large number of system components FreeBSD, GBDE, UFS2, malloc, and so on. He believes that the HTTP working group should admit defeat - and start all over again.
As an exemplary fiasco, Paul-Henning Kamp cites the example of SPDY. In classic works on project management, it is said that “the prototype system should always be discarded”, here Kamp refers to Frederick Brooks and the book “Mythical man-month or How software systems are created”. According to him, the SPDY protocol adopted as the basis for the HTTP 2.0 specifications is exactly the prototype.
SPDY was accepted even before the working group completed its previous assignment, and then spent a lot of time to bring SPDY to mind, correcting the shortcomings and errors.
')
And then everyone suddenly discovered that SPDY doesn’t even solve a lot of important problems for which they have to resort to some simplification of the HTTP concept itself.
Kamp's letter was sent in response to the acknowledgment by the chairman of the working group, Mark Nottenham, that "we may not be able to do everything correctly in HTTP 2.0, and we have not yet coped with everything," and therefore decided to "start a discussion on HTTP 3.0." Nottenham emphasizes that the working group is working on tight deadlines, in a hurry.
“Now even the chairman of the working group has publicly acknowledged that the result of the work is a partial fiasco and that we have to replace HTTP 2.0 with something better“ soon, ”says Paul-Henning. - So what exactly do we get from continuing this work? Maybe it would be better to consider the current situation with cryptography and protection of private data much deeper than to publish a protocol with a cryptographic patch that does not solve problems and interferes in many applications? ”
According to the developer, the adoption of the HTTP 2.0 standard just because the deadline has arrived and it needs to be done according to the procedure is not necessary for anyone, it only takes time away from everyone, leading to additional security risks without any significant benefit.
“Wouldn’t it be faster to start solving this problem - creating a protocol that _can_ replace HTTP 1.1 in all scenarios and really will be an improvement in _vseh_ scenarios?”, Paul-Henning Camp asks a rhetorical question. He calls SPDY to be considered an interesting prototype, which clearly showed the need to improve HTTP 1.1, but immediately set about developing a new protocol that replaces HTTP 1.1 (taking into account all the developments of SPDY).
Source: https://habr.com/ru/post/224333/
All Articles