Meeting Notes: Architecture, the Lost Years

28 views
Skip to first unread message

Jeremy W. Sherman

unread,
Sep 17, 2013, 3:51:54 PM9/17/13
to Book Club
John Gallagher, Ross Hambrick, Sean McCune, Steven Harman, and Jeremy
W. Sherman discussed the presentation “The A Word: Architecture, the
Lost Years” by Bob Martin (https://vimeo.com/30083598) recorded at a
software craftsmanship group meeting.

- What Uncle Bob proposes seems "heavier"/harder/more enterprisey.
- Very much at cross purposes to most of our tooling.
- Feels like microservice architecture in some senses – which has
its costs in deploy/versioning/etc.
- Talk of gems at the end kind of misleading - could start doing
this factorization and boundary/interactor/entity thing within a
monolithic app.
- Rails perhaps popular because it makes it easy to rapidly bake mud
pies and ignore use cases in favor of building out from the UI (only
to bite you later).

- Mobile an interesting problem arena – "models in the cloud"
- Wrap remote API to put in your model terms, so you aren't too
coupled to the backend API.
- CRUD to Rich Client hump – when leaning too hard on API to be your
"model" layer starts to hurt.
- Unreliable APIs also an issue – you want your own sane stable
model that works even when your backend doesn't

- The Boundary/Interactor thing is kind of confusing – is it app or
backend specific or what?
- See slides starting slide 11 in Keynote:
https://dl.dropbox.com/u/4730299/Architecture%20the%20lost%20years.key;
Jacobson book (Object Oriented Software Engineering: A Use Case Driven
Approach) might also help
- Two boundaries – front and back end.
- Push all the external events and persistence to the margins, and
in the middle is your network of "plain old objects" that represent
the domain and everything your app cares about.
- Parallels to ViewModel idea, but ViewModel and all of MVC really
belong in *presentation* layer outside your frontend boundary
- iOS temptation: Fat ViewControllers – a lot of that
controller-ing belongs in a model that can survive separately from
your UI or persistence solution (plists/API/sqlite/…)
- Interactors know of the app DB. Start, rollback, etc. Talks to
boundary layer that forks over entity objects as needed. (This is the
Store in BNR’s MVCS notion, but better segregated.)

Eric Evans' Domain-Driven Design likely a good follow-up read here.


Feel free to add anything I missed, or continue discussion on-list.

Thanks,
--
Jeremy W. Sherman
+1 (314) 488–0053
Software Engineer
Big Nerd Ranch

bignerdranch.com | twitter.com/bignerdranch | facebook.com/bignerdranch

Ross Hambrick

unread,
Sep 17, 2013, 6:11:31 PM9/17/13
to bnr-bo...@googlegroups.com
We forgot to talk about neutrinos :(

Sean McCune

unread,
Sep 17, 2013, 7:37:39 PM9/17/13
to Ross Hambrick, bnr-bo...@googlegroups.com
A neutrino walks into a bar and asks the bartender, "How much for a beer"? The bartender says, "For you, no charge."
--
You received this message because you are subscribed to the Google Groups "Big Nerd Ranch Book Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bnr-bookclub...@googlegroups.com.
To post to this group, send email to bnr-bo...@googlegroups.com.


--

Sean McCune, Nerd
Big Nerd Ranch

Tj Usiyan

unread,
Sep 17, 2013, 8:27:47 PM9/17/13
to bnr-bo...@googlegroups.com
I wish I could have been a part of this talk. I actually think that this talk made a lot of sense. It isn't always going to work well but there are quite a few apps that I think really _should_ receive this treatment. Turn based games are one. Many music related apps fit the bill as well. 

TJ

Graham Lee

unread,
Sep 18, 2013, 5:38:35 AM9/18/13
to Ross Hambrick, bnr-bo...@googlegroups.com
Putting myparticle-physics-graduate hat on, he did a good job of describing the history of neutrino physics accurately. It was Pauli who proposed the neutrino in 1930[*], which was detected by the Cowan-Reines experiment in 1956, and the nobel prize was awarded in 1995 to Reines. 

By the way, if any of you do a college course, make sure you get the hat.

Graham.

[*] he called it the "neutron", because Dirac hadn't yet proposed the neutron so there was nothing for the neutrino to be smaller than :)
Reply all
Reply to author
Forward
0 new messages