Combining a couple of messages: On 2004-11-01 at 07:28:14 [-0500], Matthijs Hollemans wrote: > > Should we implement our own SCM? > > No. And here is why: The goal of Haiku is to build an operating system, not > SCM tools. I am sure building your own source code revisioning system is an > interesting and fun project, and I won't stop you from building one, but it > is not a task for the Haiku developers. It is just busywork that detracts us > from the real goal. I agree. > Jobs like these don't produce the code that we need for R1. They just drain > away valuable time and resources. And anything that doesn't bring us closer > to the goal of having a working clone of BeOS R5 is simply something that > shouldn't be done, no matter how fun it is. I agree. > CVS is not perfect, but it also doesn't slow us down that much. The time we > lose because of the quirks of CVS is infintely smaller than the time we gain > by building our own CVS replacement. From a business point-of-view -- and > Haiku, Inc is a business -- it just doesn't make any sense. (Disclaimer: I > don't speak for Haiku, Inc.) I do. :-D And an SCM is definately WAY out of scope for R1. Mat Hounsell wrote: >Choice of SCM is like choice of language, editor or even colour schemes. You >can never satisfy everyone. Agreed. >If you want to talk about / plan SCM for OSBOS you could discuss it on the >Glasselevator List or write wiki pages on terminology or concepts or choices on >BeBits or write an RFC for BeUnited. Not only should you, but I encourage it. Why? Well, I think that there are a lot of potential benefits to a versioning system that people outside of developers could take advantage of. It would be great to have a document revision system built in that was better than what Word does. Likewise, something similar for picture editing. Music, DTP, etc. There are lots of different situations where you have a long running work in progress wherein it would be handy to have access to prior versions and make notes associated with each version. I think that this could/would be a great feature for some future version of Haiku. I even wanted to write it, at one point. >But IMO we should stick with what we've got unless someone provides a complete >hosting, information, viewing, build and tool set for Haiku. It is likely that we will leave SF/CVS. It is unlikely that we will do so for our own homegrown solution. Subversion or Perforce are far more likely.