Re: strings to floats/integers

[Richard D Riehle <laoXhai@ix.netcom.com>]

In article <6s1g0t$lk6@hacgate2.hac.com>,
	"David C. Hoos, Sr." <david.c.hoos.sr@ada95.com> wrote:

In response to the following message,

>
>Richard D Riehle wrote in message <6s1e9k$bgq@sjx-ixn5.ix.netcom.com>...
>>In article <35e41033.9324314@200.223.0.87>,
>
>>The conversion between strings and float and strings and integers can
>>easily be accomplished using Ada.Text_IO.   Note that for
>>
>>         Ada.Text_IO.Integer_IO
>>   and   Ada.Text_IO.Float_IO
>>
>>there are Get and Put operations defined for conversion between strings
>>and the desired numeric type.  



>True enough, Richard, but I would only use an instance of a numeric io
>generic for the case where I wanted to control the output format.  If I
will
>accept exponential notation for floating output, I would use the 'image
>attribute.  For all conversion from strings, I would use the 'value
>attribute.

>David C. Hoos, Sr.

 Formatting can often be an important issue. The attribute method,
 while a useful brute-force method of achieving the conversions,
 will not always provide the more discipline result of Get and Put.
 Also, the use of 'Value for an invalid string will raise a constraint
 error.  Would you rather see Ada.Text_IO.Data_Error.  There are many
 other reasons for using the Get and Put, but I know you are aware of
 them.  One anecdote might be useful. In converting floating point
 numbers from a VAX (the sign is in bit 15) to floating point for an
 IBM mainframe (the sign is bit 31) the Get and Put from Float_IO
 served well as a way to properly format the data from one machine to
 another.  The ability to differentiate between data errors and 
 constraint errors was important in this situation.

 Richard Riehle
 richard@adaworks.com
 http://www.adaworks.com