<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Dec 29, 2015 at 1:32 PM, Chris Muller <span dir="ltr"><<a href="mailto:asqueaker@gmail.com" target="_blank">asqueaker@gmail.com</a>></span> wrote:<br><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">But, obviously you know about ODBMS's, so probably this isn't the<br>
answer you're looking for.<br></blockquote><div><br></div><div>Well, it's the start of an answer. Sure, I know about ODBMSs, they're one solution to this problem. I can copy my data structure (objects, in the case of Squeak/Magma) to the database, then read it back later. But how does that actually work? </div><div><br></div><div>Since the database is running in a separate process, you have to serialize the objects somehow, and transmit them somehow, and when they're read back into a different image, you have to detect and accommodate missing classes and shape changes and so on. I'm sure you've implemented all this in a way that's sensible for the intended uses of Magma. If we compare your solution to say, Gemstone, I'm sure you can talk about this feature or that tradeoff that's different between the two implementations. My question is, despite those differences, are the two solutions essentially the same, or are the differences fundamental? How about if we compare with Croquet? Or CORBA? Or JSON?</div><div><br></div><div>I'm hoping this is something that people study, like oh, graph theory or category theory or parsing. </div><div><br></div><div>Thanks!</div><div><br></div><div><br></div></div></div></div>