I was wondering if there are any experienced developers out there who would care to weigh in on which of these two platforms they prefer and why. I have both and am trying to decide which is best to go with. Thanks in advance for your insight.
You need to check your requirements and then see what it will take to bring the cart to a live functional level.
If you had to add a feature that wouldn't exist on either cart what would it take to make it? I prefer to spend one day than one month to create a module. Or if you wanted to alter something else, how many files you need to mangle with to do so? 5 or 50?
If you're an adept designer/developer then I suppose you could go with either one though I've hear Magento is a resource hog/slow and I know osCommerce has very bloated code. I'd steer away from either one but there are folks here that swear by them. YMMV
Has anybody any experience with the new osCommerce Online Merchant V. 3, yet? I read somewhere that it is a complete rewrite.
I have used osCommerce 2.2 for years and have been using it because I know the source code in and out.
I tried Magento but compared to osCommerce it was just too slow. My osCommerce shop has an average loading time of 0.8 seconds according to Google webmastertools.
The major problem with osCommerce had always been that many contributions and modules required changes of core componentes - so once you actually had made some changes and implemented contributions adding yet another contribution got more and more difficult. I do not know if this has changed with the new version.
osC v3 is not finalized yet, but even if you had everything in place right now (PHP 5.3+, modules, contributions etc) is way too segmented compared to 2.2. This means it will take way longer to develop something for v3 that for 2.2. Plus the license for v3 is not ideal to attract developers.
My opinion is to keep 2.2 in place but bring it up to date for PHP compatibility as you never know what the host is going to do.
As of adding contributions and modules the thing is you need to integrate them. Integration is not copying or overriding files or modifying lines of code according to given instructions in readme files, as these assume a vanilla osc. Has to do with customizations based on your store and only.
And there isn't a framework which won't require core changes. Sooner or later hacks to the core files take place and for good reasons. In the end you need something functional and competitive. Slow is not competitive neither is segmented.
As of other commerce platforms be very careful. They use various tricks to get your attention with "open source" labels. Later on you realize extensions are not available unless you pay, but then you wonder if it's indeed open source, why someone else hasn't published the code already and the license of the core framework doesn't apply to the extensions. So best to check the license too. I prefer to go for GPL exclusively.
stocks around 2500 products listed in approximately 20,000 different categories
How on earth can you have more categories than products? on 1 to 8 scale too. I don't want to think how many problems you're going have with SEs.
Either you're nesting categories way too much, or you do not utilize the functionality of the cart. Keep a higher than 10/1 products to categories ratio and if you need the products grouped in other ways use fields, filters, zones, or other product column identifiers, for example manufacturers, so customers can filter listings if they want to and locate the product fast.
We sell consumables so one product can fit a number of different machines. The machines are categories in OSc and one product that has compatibility with multiple machines is then listed in multiple categories. I would be more specific to the products we stock to give a better example but I'm not sure WebMasterWorld allows it.
It's subject to the merchandise and marketing. I would try to generalize the machines more and move the specifics/details those than belong to rare searches, to the the product level or have them as additional options for filtering. Are all the keywords in the categories critical for people to find from SEs?
If you're exposing to SEs multiple categories (that's default for carts) showing the same products, or products listed are more or less the same, is not very good. You're also nesting from what I understand, I don't think you expose 20k category links at once, which implies the navigation can be quite deep for someone to locate a product via a categories tree.
Another way is to expand the machine properties so when these are combined to bring up a set of products assuming there are common properties. This can service better both SEO and user navigation.
I don't think it's a fair comparison to OS Commerce. OSC is just plain old. It's not a fine wine ether.
Magento is far more complex, it has extension modulation that OS Commerce lacks (in the 2. fork) It has multiple site functionality, multiple template functionality, but more is not always better like less it not always better with OS Commerce.
Magento requires more files but many are modified duplicates of the core files being overwritten. Directories are used for routing. Magento is actively updated and has a wide range of extensions You don't have to do all the programming you have to do to hack OSC as it's not secure out of the box (OSC). I know it can be hardened but it will bit you if you don't. Some assembly required.
Between OSC and Magento? no question Magento.
Sound advice about looking at other options. One size does not fit all.