random k-sample of a file

[Andrei Alexandrescu]

Denis Koroskin wrote:
> On Fri, 10 Oct 2008 16:53:29 +0400, Andrei Alexandrescu 
> <SeeWebsiteForEmail at erdani.org> wrote:
> 
>> Jarrett Billingsley wrote:
>>> On Fri, Oct 10, 2008 at 5:55 AM, Denis Koroskin <2korden at gmail.com> 
>>> wrote:
>>>> On Fri, 10 Oct 2008 08:27:37 +0400, Jarrett Billingsley
>>>> <jarrett.billingsley at gmail.com> wrote:
>>>>
>>>>> On Thu, Oct 9, 2008 at 11:28 PM, Andrei Alexandrescu
>>>>> <SeeWebsiteForEmail at erdani.org> wrote:
>>>>>
>>>>>> On a different topic, your use of std.stream reminded me. I was 
>>>>>> thinking
>>>>>> of
>>>>>> yanking it from Phobos. Thoughts?
>>>>> ..what?  You want to remove most of Phobos' IO capabilities?
>>>>>
>>>>> I'm guessing there's something else you'd like to replace it with, but
>>>>> I wonder what it could be.
>>>> Yeah, why need them when Tango is soon to be used with Phobos 
>>>> side-by-side
>>>> ;)
>>>>
>>>  That's kind of what I'm thinking.  It seems like a waste of time for
>>> _another_ IO framework to be developed when we have two already.
>>> Priorities, priorities.  How about fixing frontend bugs instead?
>>
>> Well Walter wouldn't be working on it anyway. The thing is, if we want
>> e.g. std.algorithm to work on I/O, there must be integration of I/O with
>> the range concept. As far as I can see things, ranges will be quite a
>> powerful currency allowing many std components to trade data with one
>> another.
>>
>> Andrei
>>
> 
> Great! Then, how about getting familiar with Tango IO and building your 
> concept on top of it? ;)

That would make Phobos dependent on Tango, thus running into those 
licensing issues I have little understanding of.

Andrei