[Info-vax] OS implementation languages

[Arne Vajhøj]

On 9/9/2023 8:16 PM, Johnny Billquist wrote:
> On 2023-09-09 23:44, Arne Vajhøj wrote:
>> On 9/9/2023 2:44 PM, Johnny Billquist wrote:
>>> On 2023-09-09 19:25, Arne Vajhøj wrote:
>>>> On 9/9/2023 12:53 PM, bill wrote:
>>>>> On 9/9/2023 12:14 PM, Arne Vajhøj wrote:
>>>>>> On 9/9/2023 11:45 AM, bill wrote:
>>>>>>> Nice thought, but the particular problem I was fighting was
>>>>>>> inherent to PHP and the programmer can only stop it by using
>>>>>>> a better tool.
>>>>>>
>>>>>> You are aware that PHP is Turing complete?
>>>>>
>>>>> Which means what in the concept of security?  It has nothing
>>>>> to do with the syntax or even the function of the programs
>>>>> written with it.
>>>>
>>>> It means that you did not have to rewrite in another language to
>>>> fix the problem.
>>>
>>> That definitely does not neccesarily follow.
>>>
>>> Consider for example C and some language like BASIC, which have full 
>>> control over strings.
>>> Both are turing complete. But if you want to avoid the problem of 
>>> strings being handled as pointers to chunks of memory terminated by a 
>>> NUL, then you need to move away from C (to for example BASIC).
>>> You cannot "fix" the problem of how C looks at strings.
>>>
>>> The turing complete aspect have nothing to do with that.
>>
>> I cannot follow your argument.
>>
>> Being turing complete means that it can do anything that
>> the abstract turing machine and any other turing complete
>> language can do.
>>
>> It does not mean that something is easy or best practice
>> or anything else.
>>
>> I consider C strings a bad design. But I cannot imagine
>> any problem that could not be solved with C. In most cases
>> I think there would be better languages that will make it
>> possible to implement the functionality faster and
>> with less risk of errors. But it is still possible to
>> do it in C.
> 
> If the problem is something in the construction of the language (which I 
> believe was the complaint about PHP), then the fact that it is Turing 
> complete don't make the constructs any better. Just because you can 
> solve every problem with the language don't mean that all languages are 
> equal.
> 
> If you don't believe in that, then may I suggest you start writing 
> everything in TECO. After all, it is also Turing Complete.

I get that.

In fact that was what I wrote in what you replied to.

But my discussion with Bill was not whether it was a good fit to
solve the problem but whether it was possible.

Turing complete indicate possible.

Turing complete does not indicate good fit for solving the
problem.

Millions of web sites running PHP indicate that PHP is a good
fit for solving web security problems.

Arne