X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f X-Recipient: geda-user AT delorie DOT com Date: Tue, 30 May 2017 17:22:37 +0200 (CEST) X-X-Sender: igor2 AT igor2priv To: geda-user AT delorie DOT com X-Debug: to=geda-user AT delorie DOT com from="gedau AT igor2 DOT repo DOT hu" From: gedau AT igor2 DOT repo DOT hu Subject: Re: [geda-user] [pcb-rnd] up next: subcircuits (a.k.a. footprint model redesign) In-Reply-To: Message-ID: References: User-Agent: Alpine 2.00 (DEB 1167 2008-08-23) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Reply-To: geda-user AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: geda-user AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk On Tue, 30 May 2017, Roland Lutz wrote: > On Tue, 30 May 2017, gedau AT igor2 DOT repo DOT hu wrote: >> 3rd: pad stacks and blind/buried vias (with the lowest score possible - >> seems to be a lower priority task with no real user demand, postponed for a >> later date) > > I'm not actively doing PCB layout right now and haven't ever used blind or > buried vias, so maybe my opinion shouldn't be weighed to heavily here, > however: both "professional" electronic engineers I talked to which were > considering switching to PCB for layout told me that the inability to use > blind and buried vias was the mayor show stopper for them. Sure, thanks! A year ago I made a switch in pcb-rnd "project management": I realized I shouldn't care too much about the virtual ("would-be") users because they won't ever become actual users in 99.9999% of the cases, even when you spend days or weeks implementing their feautre request. I have multiple examples on this. (You can still see a failed experiment on this at http://igor2.repo.hu/cgi-bin/pcb-rnd-poll.cgi) Instead, we favor the feature request of existing pcb-rnd users. If there are enough requests from them, requests of would-be users go at the end of the queue. This method worked out very well. When I was doing the old method, pcb-rnd was really a one-man show. Sometimes I implemented what random people requested, but they nearly never cared to test the result (requesting a feature is for free, testing software is not - see below). Since I am doing this new method, we are constantly growing. So it's more beneficial to keep those users/contributors/developers happy who are really using pcb-rnd than those who might one day consider using it (but most probably won't). Anyway, converting from virtul user to actual user is pretty simple, and I think we constantly demonstrate that we implement things in short time and that such a conversion is no waste of time. So if there's a would-be user now, who really wants a specific feature, he shall invest the time to become an actual user knowing that once that happened he can make the feature request and will probably see the feature implemented in weeks or months. I also find this fair economically: a small investment required from a would-be user before the large investment from the developers is made. Else development time becomes valueless and there would be 1000 pending featur request that then noone will ever use or even try out, wasting development time on useless features instead of focusing them on useful ones. From another aspect: if I ask you whether you want pcb-rnd to support spiral traces, why would you ever say no? Even if you think you won't ever use pcb-rnd, you have nothing to lose by saying "yes"; but if you say "no" and then you figure you want to use pcb-rnd there's one feature less. So from your seat the right answer is always "yes", because that costs you nothing and no risk involved, while a "no" would mean some marginal risk. This why I figured it doesn't make sense to ask would-be-users or non-users. So bbvias and padstacks stay low prio until there's real user demand. Even if pcb mainline implements bbvia meanwhile. (And if that happens, then please do tell us wether that made those two guys switch to pcb.) Regards, Igor2