IETF protests X-prefix in protocols
The Engineering Council of the Internet has published a new RFC6648 , which strongly recommends discarding X- type prefixes in any protocols. The new RFC is the result of discussion and consensus in the IETF community.
As is known, X- is traditionally used in the naming of experimental parameters, which allows them to be used without conflicts with other functions. This practice began in 1975, with the parameters of the FTP protocol.
For example, HTTP now uses the following experimental headers: X-Requested-With, X-Forwarded-For, X-ATT-DeviceId, X-Wap-Profile, X-Frame-Options, X-XSS-Protection, X-Content- Type-Options, X-Forwarded-Proto, X-Powered-By, X-UA-Compatible.
IETF specialists believe that in this case the negative effects are stronger than the positive effect. The problem is that experimental implementations tend to leak into an area that is covered by standards. Often, after refining the experimental function, it is necessary to maintain the support of the old X- version for backward compatibility. For example, this is what happened with the x-gzip and x-compress media types in HTTP / 1.1 .
')
Another danger is that the experimental parameter becomes generally accepted de facto bypassing the standardization procedure. A similar situation is observed with the prefix -webkit- * , when the only implementation threatens to turn into a world monopoly.
As is known, X- is traditionally used in the naming of experimental parameters, which allows them to be used without conflicts with other functions. This practice began in 1975, with the parameters of the FTP protocol.
For example, HTTP now uses the following experimental headers: X-Requested-With, X-Forwarded-For, X-ATT-DeviceId, X-Wap-Profile, X-Frame-Options, X-XSS-Protection, X-Content- Type-Options, X-Forwarded-Proto, X-Powered-By, X-UA-Compatible.
IETF specialists believe that in this case the negative effects are stronger than the positive effect. The problem is that experimental implementations tend to leak into an area that is covered by standards. Often, after refining the experimental function, it is necessary to maintain the support of the old X- version for backward compatibility. For example, this is what happened with the x-gzip and x-compress media types in HTTP / 1.1 .
')
Another danger is that the experimental parameter becomes generally accepted de facto bypassing the standardization procedure. A similar situation is observed with the prefix -webkit- * , when the only implementation threatens to turn into a world monopoly.
Source: https://habr.com/ru/post/146549/
All Articles