Re: A very quick, informal survey

• Previous message: Randy Howard: "Re: measuring job performance"
• In reply to: Steven O.: "A very quick, informal survey"
• Next in thread: Nick Landsberg: "Re: A very quick, informal survey"
• Reply: Nick Landsberg: "Re: A very quick, informal survey"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Date: Thu, 08 Apr 2004 01:00:21 GMT

In article <og49709khdbf1rr09thq6ko7vc8vpk9glm@4ax.com>, null@null.com says...
> I write documentation. The company I'm working for is developing a
> product that transmits data from one computer to another, and the
> order or sequence of data transmission is important.

Then the product is incomplete. It should handle byte-order issues
on both ends of the pipe, for all platforms it supports, period.
Looks like a faulty specification.

> We need to specify whether the data transmission and storage is
> Big-endian or Little-endian.

Actually, you need to specify a specific transmission byte order, so
that your own agents (and any others people may be able to plug into
them on various platforms) no whether the translation on the other
end is a no-op or a byte order swap. Google for "network byte order"
as a hint to how this is typically done.

> I've argued that, in addition to stating
> which it is, we need to explain what these terms mean. The
> programmers claim that any programmer worth their salt will know what
> these terms mean.

That's true, but any application worth its salt will make byte-order
issues transparent to the end user. Since you seem to be writing docs
for "salty programmers" instead of users from the paragraph above, I
assume it's some sort of SDK or library. In which case, you still
should handle byte order yourself, rather than push it on to each
user of whatever this product is. Worst case, you publish information
on the order of transmission, which should not vary (over the wire)
regardless of host platform byte order. If they try to claim
"performance issues" they're deluded and/or blowing smoke up your
backside.

In light of the above, I think you're actually asking the wrong questions
here.

-- 
Randy Howard                
2reply remove FOOBAR 
      

• Previous message: Randy Howard: "Re: measuring job performance"
• In reply to: Steven O.: "A very quick, informal survey"
• Next in thread: Nick Landsberg: "Re: A very quick, informal survey"
• Reply: Nick Landsberg: "Re: A very quick, informal survey"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]