Moved from ops#2 (closed)

Moved to www since we only keep ops for meetings.

Added PR label

There's a rewrite rule pointing at the Eagle files. It should probably point at the new KiCAD files:

rewrite ^/schematics http://neo900.org/stuff/eaglefiles/proto_v2/2_2014-12-21/GTA04b7.sch.pdf redirect;

mentioned in issue ops#4 (closed)

mentioned in issue ops#5 (closed)

More bullets from @wpwrak

"kicad arcs are defined by center and two angles. however, the direction can be either clockwide or counter-clockwise. this ambiguity is resolved by limiting arcs to < 180 deg." "the converter doesn't know this, and tries to make arcs >= 180 deg, instead of breaking them up into two arcs, each < 180 deg. this is what caused the weirdness in the sensors page." "arcs are so much fun. there are lots of ways to define them, and to resolve ambiguities, and no two pieces of software use exactly the same conventions :)"

@joerg_rw : " we should contrinute bugfixes to eagle2kicad"

So, I'm writing in the 0021 branch

Posted some comments on the Tue Aug 2 version on #pr. Here they are, slightly cleaned up, so that we don't miss them:

• I wouldn't say we "build upon" the Openmoko phones. For separating the modem, we can cite as references, besides Neo900: Openmoko Neo 1973 and Neo FreeRunner; Golden Delicious GTA04; and maybe the upcoming Pyra.

• that's why I'd also not say "plagues all" but "plagues most". Also, there's always an exception you don't know about ;)

• as far as I remember, you couldn't detach the modem from power in the Neo 1973. The Neo FreeRunner had this functionality, though.

• Just separating the modem doesn't mean you can monitor all this. You need circuits specifically designed for the purpose.

  The absence of RAM sharing does apply to all these devices, though. (So this part is correct).

• "won't affect Neo900" still seems a bit too strong, since it can affect the modem.

One item for the Eagle section:

Another welcome result of the migration is that the design files are now hosted in the project's public git repository: https://neo900.org/git/?p=ee;a=tree

"Hosted" may sound a bit odd. The word occurred to me because the underlying change is that the design files were before maintained by Nik internally (apparently without version control whatsoever), and the mailed us copies every now and then, which were kept either internally or were somewhat haphazardly placed in various placed on neo900.org. Now we've come out of the stone age, and all this has a well-defined location with proper revision control. So it's not only a transition from stray files to git, but a more profound change of domain and workflow.

Maybe also briefly mention the server migration:

Last week, we also migrated neo900.org (including the shop at my.neo900.org) to a new server. The old server was running low on disk space, and, when Neo900 has been being mentioned in the news, it has been too weak to keep up with the resulting flood of Web site visits. (I.e., we have been "slashdotted" a number of times.)

The migration went fairly smoothly and everything should be up and running again. If you spot any problems, please report them to contact@neo900.org

---

For "slashdotted", maybe pick one of these links: https://en.wikipedia.org/wiki/Slashdot_effect http://www.urbandictionary.com/define.php?term=slashdotted

Regarding "won't affect Neo900" being too strong, it seems that we can safely say that, with the exception of the inside of the modem, we don't have ASN.1 anywhere where the implementation couldn't be fixed if broken. (ASN.1 is used in many places, e.g., also in GnuTLS, e.g., for X.509 certificates, which then leads to many, many dependencies [1].)

[1] For a pessimistic estimate, run this on a debianite (this may take a while): apt-rdepends --reverse libtasn1-6 | grep 'Reverse Depends' | grep -v ' lib' | sort | uniq

almost there... Split ASN.1 vulnerability to #9 (closed) .

Remaining: more details about the whitepapers.

Notes from the backlog:

@wpwrak

"Full package for schematic and board design, etc. Design rule checking. User-defined symbols and footprints. Gerber/ drill file creation. Sophisticated graphic interface. Active user community"

does that come from wikipedia?

and the comparison chart is a bit obsolete ;) kicad now has some simulation

so don't trust wikipedia too much :]

@joerg_rw makes some remarks suggesting not all is positive in the migration to KiCad, which is a great point for a balanced review.

The dialogue makes for a good review of things one wants to do and can't, and how to go around them or actually do them "the KiCad way". Some review needed to ensure the "KiCad tips" section is useful.

the "python" tag in "scripting support" col is pretty much incorrect actually kicad has not even a forward/back-annotation between schematics and layout you could script network highlighting and forward/back annotations, IF kicad knew any scripts in eeschema, and eeschema would know highlighting of multiple objects pcbnew layout editr has a "Scripting console" that talks python, but it seems no docs for that thing are available so you only can guess how to use it, and you can't load and execute arbitrary python scripts in a simple and easy way

a click on a component in schematics and then say "show me this thing in layout" is basically completely impossible, you need to do it manually the other way around might work with xdotool and lots of hackery

Actually you need to open bothpcbnewandeeschemaviews so they can talk to each other. When the zone under focus is too crowded, the result can lack precision, so better use the zoom feature to adjust the view to your desired component.

it seems it has no network highlighting either.pcbnewhighlights traces, buteeschemaonly highligts wire (segments) but no networks and more often than not the both ends of a component are indistinguishable pcbnew will also highlight nets (without involving eeschema), but that happens typically when you add a new trace to

UI quirk:

dang, when the component I select in pcbnew is visible on the eeschema window, then it does not center but only place that haircross there, which jumps back to mouse pointer as soon as mouse moves which means you must NOT move mouse at all and hope you find the flimsy haircross before it nevertheless jumped back to cursor

you definitely want two screens. with one it's just agony if you can add a 3rd screen, even better: #1 (closed) schematics, #2 (closed) pcbnew, #3 data sheet or communication meh, the status line gives it away, not that this helps a lot

White papers

http://talk.maemo.org/showthread.php?t=93498

mentioned in issue ops#6 (closed)

does that come from wikipedia?

Yes. Seems that the KiCad marketing department was busy :)

Suggesting the following minor edits:

• s/and comes/and come/
• s/finishing touch/finishing touches/
• s/has the same flaw as Eagle/shares a major flaw with Eagle/
• s/can now operate/can operate/ # two "now" in a row are poor style
• s/We found out that // # we knew that for a good while :)
• "improvement over that." maybe s/that/the corresponding Eagle experience/ ?

Awaiting @joerg_rw 's ack or nak. (Please use the OK tag to confirm what's in branch 0021 is ready for publication, or comment with required changes.)

Reassigned to @joerg_rw

mentioned in issue ops#7 (closed)

WHICH OK TAG? http://wstaw.org/m/2016/08/09/plasma-desktopnQ2277.png

also which text is this all about?

https://git.devuan.org/neo900/www/blob/0021/content/0021-migrating-away-from-eagle-to-kicad.html

ok, I finally figured what's meant by "use the OK tag" - sorry I didn't get tags are labels and I need to edit them

Added OK label

for future similar text production: please add a link like

hellekin @hellekin commented about 2 hours ago https://git.devuan.org/neo900/www/blob/0021/content/0021-migrating-away-from-eagle-to-kicad.html

on top of the issue so we can find and use it (if it's not possible to keep the text inside the issue).

And please use markup or plain ascii for the draft, any HTML meta can and should get added only after the content got finalized. Many thanks

Removed OK label

hellekin @hellekin about 7 hours ago Removed OK label

??? Hmmm, no...

Added OK label

Status changed to closed

Removed PR label