Hi there!
Well first read this
https://www.eevblog.com/forum/product-reviews-photos-and-discussion/review-of-owon-sds7102/ thread, if you haven't already. It is long, but it has plenty of details as well as some common problems with the scopes such as the grinding fan problem or the channel 1 problem (which should have been solved for some time now).
As for the type of scope itself, you should spend some time and try to think how you're going to use it. What you wrote as intended use is a bit vague. You'll need to "assign" priorities to the uses which you consider more important than other and be aware of the limitations of each version of the scope. For example the sample rate listed is for one channel use only and if you want to use both channels, the sample rate will be cut in half. The result of that is that a two 100 MHz signals viewed by each probe won't be as accurately shown as one 100 MHz signal viewed by one probe. For PICs, AVRs and some CPLDs that may not be a problem, but some FPGAs are fast and you may need more bandwidth and sample rate in order to view their signals correctly.
Next, it seems that you're thinking if you should buy a logic analyzed or a scope. If you're going to get seriously working with electronics, you'll need both!
Why? Well the SDS7102 does have large sample memory (10 Mpoints per channel), but it's a scope and it doesn't do any protocol decoding or similar. What that means in practice is that if you have a device using say I^2C device and you're probing the data line, you'll need to count ones and zerores by hand and then decode convert them manually into hex digits and then figure out what's happening. Decent logic analyzers should be capable of doing that automatically. Good ones (in the low price range) will even be able to decode some parts of the protocol and tell you for example which parts of the message are the actual data being transmitted and which are just ACKs/NACKs and other protocol details. If you get a scope you'll have to do all that by hand and it gets tedious quickly. I know that because I decided to get an SDS7102 instead of a logic analyzer. All is not horrible though, since if you have a working bus, you won't need much details about the shape of the digital signal itself so you'll be able make long captures by increasing the timebase. If you're good with programming, you could import the data captured by the scope as a .csv file and then process it to do some decoding on a PC, if you want that. Still, a logic analyzer is more comfortable for debugging some parts of digital logic.
On the other hand, if you're designing or troubleshooting an analog circuit (or a digital one where you expect problems with signal integrity), you'll definitely need a scope. A logic analyzer simply can't replace a scope in such usage scenario. Scope will give you the ability so actually see the waveforms and based on that you'll be able to decide what needs to be done. If you're going to be repairing modern switch mode power supplies, you'll need a scope to solve more complex problems.
Also do note that the 200 MHz scope called SDS8102 is in some areas no better that the 100 MHz SDS7102 and that areas may be important to you. The scope has bandwidth of 200 MHz and sample rate of 2 GS/s. If you use both channels, you get 1 GS/s per channel, so each channel is as good as whole SDS7102. In that mode, each channel can store up to just 10K points of data! To get the 10 M points, you'll need to run both channels in sample rate of 500 MS/s or lower. One channel running in 1 GS/s mode can fill all 10 M points, but if you run it in 2 GS/s mode (which may be needed by some very fast FPGAs), you'll get again just 10 K points.
As far as Owon is concerned, their hardware, see what people are talking about in the thread I linked to. My impression is that their hardware is not that bad, but the software side isn't very good. The PC software is horrible and the scope could do with a bit more and a bit better implemented software features.