Continue With The Operation, II
So, here we are.
As you've probably noticed, this storefront has changed significantly. The thing is that Shopify, which I used to run my corporate Web presence since starting the business, changed its admin interface in such a way that it would no longer work with my browser. That was a dealbreaker and it meant I had to find a new home for North Coast online. Ironically, it was not Shopify's failure to adequately address GDPR issues that drove me away, even though I was angry about that. The admin-interface thing is mission-critical and has no permanent workaround, and being forced to move because of that means I no longer need to have a stake in Shopify's GDPR situation. Since I've managed to complete the move to a new host before May Fool's Day, I can focus any GDPR efforts on just the new system. (Also, upon reading the actual GDPR regulation itself instead of summaries from people with a stake in selling compliance-consulting services, it appears that GDPR's actual applicability to me, and the things I must to do meet the obligations that do apply to me, are more limited than I'd feared.) I'm just lucky that I was able to limp along with Shopify in an alternate browser long enough to evaluate alternatives, save the data, and so on, instead of having had my shop down throughout the transition period.
This new store is hosted by BigCommerce. It's another turnkey Web storefront service, with a general selection of features comparable to those of Shopify. Some advantages, some disadvantages; at least for the moment it doesn't have the dealbreaking issue of not working in my browser. I posted an article similar to this one on Shopify last week but am not bringing it over to the new system because it's largely obsolete. Here, instead, is a rundown of different issues relevant to the move updated with what I know now, having completed most of the move.
Look and feel: After more than a year of customization (including a significant period working on the Web site before I opened it in August 2017) I had the Shopify storefront looking pretty much as I wanted it, and ideally, I'd be able to use the same Web design on the new system. Unfortunately, that was not really possible. Themes from one provider are not portable to another because of different data models (as well as likely licensing problems) so I had to just buy a commercial theme for the new place that came as close as possible (still not really close at all) to the existing Web design, and then I tried to customize it. One of the first things I did was go through and remove all the dependencies on Google. Your browser may still hit Google-controlled servers occasionally in the case of embedded YouTube videos, but the privacy exposure is at least much reduced. It's quite gratifying to open the front page of this site with my browser's "Network" tool and see no hits to any servers except my own and BigCommerce's. Getting all the design right will be an ongoing process probably taking months. The new site is now at a level I consider basically usable, but it's still not where I want it. On the plus side, the new system provides much better tools than Shopify did for customizing themes. In particular, I can now make and test my theme edits on a local desktop machine, with some approximation of proper version control, instead of having to use the provider's servers for everything. On the minus side, those tools are based on NodeJS.
This Web log: I've been able to move all the entries in this Web log that I wanted to move. I still need to keep my eyes open for broken links, because I moved all the images over from Shopify's CDN to BigCommerce's and if I missed any URLs then they will eventually break. The canonical URLs for entries (which is where you would, ideally, bookmark them) have changed but I think I have got redirects in place for all entries, so links should not break. However, existing comments on entries are no longer displayed, and at least initially, it won't be possible to comment on entries anymore. The new system does not offer native commenting, only a gateway to Disqus, and I don't want to use Disqus because of its ties to Facebook and Google accounts. Even if I didn't personally object to that kind of third-party login system, using it would create a bunch of extra GDPR obligations on me. (So does commenting at all, though less so.) So the only practical way I can see to have a commenting system here in the future would be if I can host it myself to avoid involving third parties I don't trust and you shouldn't, and although self-hosted commenting may be possible some day, it's pretty far down the priority list. I'm going to save the existing comments and might possibly be able to restore them if I re-introduce commenting here in the future.
RSS: Unfortunately, BigCommerce just doesn't do RSS. They used to, and they ended support for it without explanation of why. I am a bit annoyed by this because when I was evaluating BigCommerce as a possible host I saw the information (which is still posted prominently) about how to set up RSS feeds, and didn't see the note that it does not actually work anymore, so I feel that they misled me. Also, they do still provide RSS feeds from shops that existed before last Fall, so it's hard to claim that they have a hard technical limitation preventing them from providing this service to me too. However, it's not a big enough issue to be worth aborting the move from Shopify to BigCommerce and looking for yet another replacement system. If you are subscribed to the RSS feed for this Web log, you are now getting a static hand-written RSS file that just warns you the feed isn't working. I'm hopeful that I can write some custom code of my own to generate RSS feeds, and if so, I'll be able to make them better than the Shopify RSS feeds used to be (in particular, Shopify's RSS file was a half megabyte, which is insane, because it contained too many items and the entire text of most of them). But I cannot promise any specific schedule for when this will be working.
Other URLs: URLs pointing to specific products should at least have redirects (so your links won't break), and may not change at all. URLs pointing to specific variants in the old shop will probably redirect to the main page for the product in question now. Per-varient URLs are complicated by the fact that BigCommerce has a different data model; what Shopify calls "variants" have less independent existence under BigCommerce's model, with more emphasis on "options" attached to a single product.
Mailing list and user accounts: If you have subscribed to the North Coast mailing list on the old system, I still have your subscription recorded, but I'm having issues with getting it stored on BigCommerce's systems. I discovered, just within the last hour as I was preparing to send out an email announcement, that their system for keeping track of customer newsletter subscriptions does not work for me, and I'll probably have to just turn off all subscription-related features and do a self-hosted replacement. I'll send an email to the list soon describing the current situation in more detail, but it will probably change later. User accounts have come over in the switch, but passwords have not. Since Shopify records only hashed versions of passwords (information to recognize when you enter the right password, without actually storing the password itself), and BigCommerce does the same, and neither of them lets me have access to even the hashed version, there is no realistic way for me to re-create your account on the new store with your old password. You have to do that yourself. To do it, try logging in with your email address and your old (or, technically, any) password. It will reject the login and send you a "password reset" email message which you can use to set your password on the new system. You can't create a completely new account with the email address of an old one - it will reject that as a duplicate.
Discount codes: Existing discount codes that were in effect before the move, remain valid, or at least they're supposed to. Contact me if you have trouble. One piece of custom software I was using for issuing new discount codes at random to site visitors, broke and will be unable to issue any new codes until I re-write it.
Multi-currency support: It's still the case that final charges to your credit card, and our binding price offers, are in Canadian dollars only, but (in a change from the previous arrangement), now the price in your chosen currency is shown prominently on every page of the checkout process, and I'm hoping that will cut down on the confusion when people add to the cart using the USD price, see the same price in CAD (same amount of actual money but written with bigger numbers) on the final checkout page, and mistakenly blame me for hitting them with an unannounced price "increase."
Cookie warning: because the European Union needs to prove that they are "getting tough" on Mark Zuckerberg, I've been required to implement a cookie warning. It's an annoying bit of theatre and will probably cost me some sales, but that's how the Web works in 2018. What lesson do you think the Zuck will learn from it?
Backend server: I've moved my backend server (which hosts some features that aren't easy to host on the e-commerce provider's site, such as my own analytics [I don't want to use Google Analytics or similar because of the privacy nightmare], large audio files, and so on) from a cloud host in Japan to Germany. I'd really prefer to keep it in Canada, but the service I want to use doesn't have a site here. So my thinking was: USA - out because run by an unstable regime which claims absurd snooping powers, not that you can really avoid US contact when using a credit card anyway but at least I can keep my own activities outside their borders; Japan - my own first choice given I can't have Canada, because they're reasonably stable and they don't really care what foreigners do anyway, so there's unlikely to be much government interception of my and your data, but Japan doesn't have trusted status under GDPR; Singapore - doesn't have GDPR status and I trust the government there less than the Japanese government; England - out of the question because of the "Snooper's Charter" (and when that came in, which was before this company existed, I already went through a painful process of pulling all my other projects out of the UK), also some stability questions related to their exiting the European Union; and that leaves Germany - which I'd been trying to avoid pre-GDPR because I didn't want to be covered by their local nonsense, but since GDPR basically is the German local nonsense scaled up to cover all of Europe, it's no longer something I can avoid, and at least putting a server there means there's no question they will trust I have it in a secure country. Anyway, the backend server move did not go as well as I'd planned and I had a total of about 12 hours of down time on that server (in a couple of separate incidents) during which analytics data was not being collected and some Web site features didn't work. If that hit you, sorry! It was supposed to go a lot more smoothly. But at least it's over now.
Wishlists and gift certificates: I think we have wish lists and gift certificates now. I'm not sure yet exactly how they work. They came included with the BigCommerce plan and I haven't really explored them because I've had other priorities. But if these features interest you, you can go try them out.
I'm pleased to announce that SchneidersLaden, in Berlin, is now carrying the Leapfrog Filter. (see also English version) If you're in or near Berlin, you can drop by their showroom and try out the module live. The SchneidersLaden initial order was big enough to require a Canadian export declaration (not needed for the smaller packages I send when individuals order modules one or two at a time) and getting the proper authorization for that was an exciting adventure which maybe I'll write about another time. As a reminder, most North Coast Synthesis products are also available from synthCube in the USA. If you're in Europe or the USA you may find it easier to deal with these resellers than to try to order from me directly, though the option of ordering direct from my Web storefront remains, and ordering from me directly does help me come closer to breaking even.
I've had to put development of the MSK 012 Transistor ADSR on the back burner while resolving these other matters. Just yesterday I heard from a reviewer to whom I'd sent a review kit - he was eager to build it even though I haven't written the instructions yet, so I had to quickly make a list of tips and pitfalls, and what I've heard back was that he liked the module but wanted the trigger threshold to be higher, for better compatibility with poorly-behaved trigger-generating modules that don't bring their outputs all the way to 0V. I'm going to have to go back and test that myself. Depending on how big I think the issue is I may end up changing the official design; mentioning it in the manual as a possible "mod"; or doing nothing. Raising the threshold doesn't come without cost because it means other poorly-behaved modules, with low "high" trigger voltages, won't be able to trigger the envelope at all; so I have to trade off what kinds of misbehaviour I expect to really see in the field.
◀ PREV Modular synthesis intro, part 13: Envelope generators || Modular synthesis intro, part 14: Multiples NEXT ▶