Keep all books away from them other than the core.
Make flash cards with the Art Combinations they can cast.
Make a card for spells or a card per spell. Photocopy the character sheets in the back of ArM.
If you can as storyteller get into the Latin as much as possible for immersion. On the other hand do not be pedantic about it, not about history. If players feel the system is overwhelming and/or the historical aspects are overwhelming, they will lose interest.
I've found that, especially for new players, calculating spell levels grinds the game to a halt. When someone wants to spont a spell, you as the story guide should just estimate a level off of the top of your head and keep the game rolling (like you did with 2nd edition). Leave the guidelines for downtime.
I'm running a game at the moment for three players: one is experienced with the background but not with the system, the other two are completely new to Ars Magica.
In hindsight, I think I made a slight mistake in being too 'helpful' in calculating spontaneous magic results, etc. The system can be a bit daunting and I tried to take much of the calculations on myself, leaving the players free to roleplay, but I've probably gone a little too far on that and should make sure that they learn the numbers themselves. To that end, I produced a four-page document that breaks down spellcasting in a (hopefully) intuitive way, lists the formulae and guides them to making the calculations (and decisions on if they want to risk that spontaneous spell) for themselves.
The game has been a fairly slow-paced one so far (exploration, discussion, not much peril) and we've not had a full-blown combat yet. Depending upon what your players enjoy, it might be worth doing something similar and introducing opportunities for spellcasting as the story goes on.
What I have found is that they frequently want to know: Am I in an aura? Is there vis here? Is that thing magic, etc? So I have recently pinned down exactly the casting totals needed for doing that spontaneously (as none of them took those basic spells) so that they have more of an idea now of what they are capable at these low levels.
A thing that I did with a group of mostly non- ars players one evening was to have a laboratory competition. Each person had a magus in the same covenant and we advanced the magi three or so years at a time with all magi having the same goal.
IIRC the goals we used were (in order)
Make the covenant look the most impressive to someone approaching by the road
Find the most hilarious way to deal with a sleeping dragon and its clutch of 5 eggs that are due to hatch in 3 years and wake it.
There is a complete solar eclipse coming in 3 years. Be able to host a party in fact the best party that moves with the path of total extinction for at least 4 hours.
After each contest the results were presented and everyone voted for the best creation (it had to be someone else's). We then drank a toast to the winner.
This got people using the lab rules and spell and item creation rules pretty well.
An idea I remember someone mentioning a while ago on the forums, that I really liked, was have everyone start by making a grog character, then run a short adventure of the grogs going to the covenant, whether returning there from a mission, or being sent there from another covenant as a sign of support. Then the players can make their characters (companion, magus, or both), while the storyguide fleshes out the covenant site.
My game, when I was introducing new people, I made sure to give everyone a taste of faerie, a taste of infernal, a taste of GLORIOUS MAGIC, and a bit of the power of the divine (via dominion auras and congregations) via early adventures. An established covenant that just went through a catastrophe or rebuilding a dead site are both really strong starts for new players.
It took a while for my players to get used to spontaneous casting totals, because they kept forgetting to halve the numbers, or forgot that penetration is based off of the halved number. Some forgot to include touch range for Creo spells... But they've become pretty good at that. One player made a spreadsheet that auto-calculates spell totals.
I found the hardest part was getting the characters invested in the grogs. I helped the game get started by designing all the grogs, so none of the players felt invested in playing or keeping them around.