On Wed, Jul 25, 2012 at 5:11 PM, Lester Caine <les...@lsces.co.uk> wrote:
> Andrew Faulds wrote: > >> I top you 20 years with 37 years. I was programming in Algol in 1975 ( at >>> Warwick university ). I'm not a programmer, I'm a hardware engineer who >>> has to >>> program to make systems work. I added PHP 12 years ago to create web >>> based >>> applications to augment c and Pascal based applications. Many of the >>> concepts >>> being added make sense only as extensions to the core code, and don't >>> need to >>> be forced into general use. Adding tools that have very specialist use >>> should >>> be done as options, which we can leave out if we want to. The it needs >>> to be >>> justified switching something on by default. I have no objection to 'new >>> facilities', but only if I can also switch them off ... >>> >>> Eh, what? "Switch them off"? >> >> What on earth do you mean? "use noGenerators" and then break tons of >> third-party >> code relying on it? Or do you think you're forced to use them? >> > > The third party code would have to justify using them in my book. But as > long as I KNOW what a library is using then I have the option simply to > avoid it. At the moment how many third party codes are 'strict compliant' > and can be used safely with PHP5.4? In my book this is exactly the same > problem. For many years all of my PHP5 code played nicely with PHP4 > installations even though I never used PHP4. Nowadays you are saying "You > can't use this until your ISP updates" - so libraries NEED to manage this > or specifically avoid features that are not in general use 'in the wild' > ... ISP are only just updating to PHP5.3 since PHP5.4 is not generally > available in the core ISP distributions ... It's additions like this that > are delaying adoption rather than enhancing it. some/most of the E_STRICT messages are telling you about (possible) problems in your code. you can decide to ignore them, but imo it is a good thing that we tell those to the developer. the PHP4->PHP5 transition was pretty major, I'm surprised that your "php5 code" worked on php4 just fine. my hunch is that you only used a small subset of the language (and you didn't used oop at all, else you would remember the php4->php5 migration PITA), so you simply didn't noticed that back then. PHP 5.3 is still supported, and 5.4 is pretty new (in terms of distribution life-cycles), so imo it is perfectly fine that most distro still uses 5.3 as the latest available version. not fixing the outstanding issues or implementing enhancments/extensions which makes the majority of the php devs lifes easier also can be a problem in the long run. of course we have to find the balance, and only accept/implement reasonable changes, which are worth the BC brakes/bloat by a big margin. > Until one gets thrown into libraries that use 'short array syntax' > exclusively and have to work out what the heck is going on. We don't get > forced to USE it, but we do get forced to ensure that we understand exactly > what it is doing even if we don't and it adds unnecessarily to the library > of functions we need to understand :( I don't understand this reasoning. You said that you are ok with the new features, but you should be able to turn those off. If you have libraries which depend on the new feature, you are screwed anyway whether you can disable it or not. So you can either say that you don't want any change, or you should accept the fact that with change, there is a chance that either directly or indirectly you will be using that feature. -- Ferenc Kovács @Tyr43l - http://tyrael.hu
