Re: [icnrg] Comments on reflexive draft

"David R. Oran" <daveoran@orandom.net> Wed, 08 April 2020 15:47 UTC

From: "David R. Oran" <daveoran@orandom.net>
To: Ken Calvert <calvert@netlab.uky.edu>
Cc: Dirk Kutscher <dku@dkutscher.net>, icnrg@irtf.org
Date: Wed, 08 Apr 2020 11:47:15 -0400
Message-ID: <65E13600-7D33-4E92-A076-D580504CB927@orandom.net>
In-Reply-To: <34552F76-578B-43C6-BBF3-54930DD2DE1F@netlab.uky.edu>
References: <34552F76-578B-43C6-BBF3-54930DD2DE1F@netlab.uky.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"; format="flowed"
Content-Transfer-Encoding: 8bit
Archived-At: <https://mailarchive.ietf.org/arch/msg/icnrg/XLsiSjrNXV4O2MdEGvsbezA1I0U>
Subject: Re: [icnrg] Comments on reflexive draft
Precedence: list

On 8 Apr 2020, at 11:16, Ken Calvert wrote:

> Gents - Overall it reads really well and does a great motivating the 
> proposal.  And obviously feel free to ignore any suggestions.
>
> 1. Introduction
>
> 6th paragraph: "built on top of on reliable" [delete 2nd "on"]
Fixed.

> 3. Overview of the Reflexive Forwarding Design
>
> Bullet 1. ("The Reflexive Interest needs..."): "This name is what the 
> originating consumer will use to match against the Data object ... it 
> wishes the producer to fetch..."  I found this construction a bit 
> confusing; suggest something like "this name (or an extension) will 
> refer to the Data object(s) the originating consumer prepares to send 
> to the producer..."  That also avoids the anthropomorphism ("wishes").
>
Got it, although I prefer to avoid using the term “send” since it 
introduces a push-like mental model. How about
“This name is what the originating consumer will use to match against 
the Data object (or objects - more on this later) that the producer may 
request by issuing the Reflexive Interest.”

> Bullet 2. ("There has to be a FIB..."):  Suggest "At each forwarder 
> along the path from producer to consumer, a FIB entry must match 
> (LNPM) this name...".
Close - how about “At each forwarder along the inverse path from 
producer to consumer, a FIB entry must be present that matches this name 
via LNPM,…”.

> Also "deign" -> design, later in the paragraph.
Got it.

>
> Bullet 3. ("There has to be coupling...")  "...this accomplished by 
> the way..." -> this is accomplished by the way
>
Fixed.

> 4. Naming of Reflexive Interests
>
> 1st paragraph: "...enough information in their initial Interest" -> 
> enough information in its initial Interest
Fixed.

> Also, "...enough state must be provided by the consumer..." - I found 
> this a bit confusing.  The forwarder creates the state, from info 
> provided by the consumer.  Suggest "enough information must be 
> available for the forwarders to construct a FIB entry...", or just 
> change "state" to "information".
>
I made the simpler change of “state” to “information”

> [I have some additional thoughts about the content of this section, 
> but will send them separately.]
>
Looking forward to it!

> 7.1 Achieving Remote Method ...
>
> 2nd paragraph: "computating" -> computing
>
Fixed.

> 8.1.2 Interactions with Interest Lifetime
>
> 4th paragraph: "...a reflexive Interest arrives from upstream matching 
> the..." - suggest deleting "from upstream", it's not clearly defined 
> and doesn't seem necessary.
>
Ok, deleted.

> last paragraph: "While the above approach..." - this sentence is 
> garbled, the first clause seems to be missing a predicate.
>
yes, added “…avoids the timeout problem,”

> last sentence: "Interest-Datsa exchange" -> Interest-Data exchange
>
Fixed.

> Cheers,
>
> Ken
These will show up in -01, which we’ll issue in a few days after 
waiting to see what other comments surface form other ICNRG’ers.


DaveO

[icnrg] Comments on reflexive draft Ken Calvert
Re: [icnrg] Comments on reflexive draft David R. Oran