>>Newer, more compliant browsers, will in time not support the older tags and code
HTML is still a clearly-defined standard. Lots of manuals are on old install CDs; lots of pages, documentation, and other information is archived on CDs in everyone's drawers and shelves. HTML can't just be dropped like it never mattered.
I am repeatedly disturbed by the attitude that backward compatibility is unimportant, or useless, or wasteful. Much of our technical civilization is built with the assumption that things continue to work, and we expect them to work even after long periods - how many people actually checked their air conditioners this spring before just turning them on?
Would anyone accept a statement like "My new calculator program won't work on checking your statements from three years ago, because those are old numbers"? Or "You can't speak Latin, Greek, or Hebrew on this new cellphone, because those are ancient languages"? Isn't that the kind of complaint Slashdotters always make about M$ office programs changing document and spreadsheet formats?
New versions of computer languages should compile old programs. New browsers should display old documents. Old stuff may run (or be processed) less efficiently, but it should do what it used to do.
Re:New language versions break code? Unforgivable!
on
Perl Medic
·
· Score: 1
I have no problem with 101 ways do do the same thing, or even weird idioms; that's true in most computer languages, as it is in English and other natural languages. As John Dikstra said, you can program Fortran in any language. I *do* insist that when a new interpreter version is released, all 101 should still do the same thing(s) they used to do rather than doing anywhere from 1 to 101 different things.
New language versions break code? Unforgivable!
on
Perl Medic
·
· Score: 1
Why isn't everyone totally unforgiving of the idea that new version releases break existing products? If it's suppposed to be the same language, it's supposed to WORK. Yes, you can "deprecate" things, and suggest that they not be used any more, and even let them be less efficient, but legacy code is not supposed to be broken by the compiler, any more than running the same old code on a newer computer should break. I've been in programming since the IBM mainframe days, and this would have been considered a hanging offense - certainly lawsuit material. Imagine if the phone company decided to change its interfaces overnight and your phone stopped working - that's how ticked everyone should be.
Sadly, none of the media players I have could handle it. I thought I saw a mention of the Dirac codec in the previous/. item; is a compiled Windows version available anywhere, or do I have to build-from-source from sourceforge? (a ridiculous concept, by the way - why should every single machine have to do the work over?)
Slashdot Mirror
>>Newer, more compliant browsers, will in time not support the older tags and code
HTML is still a clearly-defined standard. Lots of manuals are on old install CDs; lots of pages, documentation, and other information is archived on CDs in everyone's drawers and shelves. HTML can't just be dropped like it never mattered.
I am repeatedly disturbed by the attitude that backward compatibility is unimportant, or useless, or wasteful. Much of our technical civilization is built with the assumption that things continue to work, and we expect them to work even after long periods - how many people actually checked their air conditioners this spring before just turning them on?
Would anyone accept a statement like "My new calculator program won't work on checking your statements from three years ago, because those are old numbers"? Or "You can't speak Latin, Greek, or Hebrew on this new cellphone, because those are ancient languages"? Isn't that the kind of complaint Slashdotters always make about M$ office programs changing document and spreadsheet formats?
New versions of computer languages should compile old programs. New browsers should display old documents. Old stuff may run (or be processed) less efficiently, but it should do what it used to do.
I have no problem with 101 ways do do the same thing, or even weird idioms; that's true in most computer languages, as it is in English and other natural languages. As John Dikstra said, you can program Fortran in any language. I *do* insist that when a new interpreter version is released, all 101 should still do the same thing(s) they used to do rather than doing anywhere from 1 to 101 different things.
Why isn't everyone totally unforgiving of the idea that new version releases break existing products? If it's suppposed to be the same language, it's supposed to WORK. Yes, you can "deprecate" things, and suggest that they not be used any more, and even let them be less efficient, but legacy code is not supposed to be broken by the compiler, any more than running the same old code on a newer computer should break. I've been in programming since the IBM mainframe days, and this would have been considered a hanging offense - certainly lawsuit material. Imagine if the phone company decided to change its interfaces overnight and your phone stopped working - that's how ticked everyone should be.
Sadly, none of the media players I have could handle it. I thought I saw a mention of the Dirac codec in the previous /. item; is a compiled Windows version available anywhere, or do I have to build-from-source from sourceforge? (a ridiculous concept, by the way - why should every single machine have to do the work over?)