Their primary difference is in their intended uses. Exploratory is most appropriate when you do not know the underlying factor structure of a data set and wish to determine it. Confirmatory is most appropriate when you think you know the underlying factor structure and you want to confirm (or refute) it. Another way of looking at it is EFA is most useful to help establish theoretical structures and CFA is most useful in confirming (or refuting) theoretical structures.
James E. McLean covers the differences nicely. On the software, while I am sure other commentators will argue for their favs like M-Plus, different R packages like lavaan, SPSS, etc., I'll offer some reasons that may help you choose one of the other, in priority order.
1. Can the software do what you want? You don't provide much detail about the specific model or problem characteristics, but M-Plus, R, Python, are pretty general.
2. What others in your department use and what is taught there. If you do not do a lot of these it will be useful to have someone to ask in person.
3. Software easy of use, including downloading and price. If you are just doing one CFA/EFA, don't know much about stats, any of these will be a hurdle and I recommend getting a statistician to do it. If you are wanting to learn, then ease of use is less important. On price, lots are free, but even if the software is expensive, like SPSS or SAS, if it is taught in your department there is probably a site license.
Exploratory is used to structure your measured variables into different domains while confirmatory factor analysis is used to check the standard tool variable is different setting to know if they fit in the same domain or differently. Well, you can use SPSS, R etc. I agree to James and Daniel view.
EFA is working on dimension reduction and reveals the items are in the questionnaire load together. Items in a construct must load together otherwise, that item or items may measure different ideas which you did not intend. On the other hand, CFA is working on confirming the constructs measured. I recommend SPSS for such analysis.
EFA is working on dimension reduction and reveals the items are in the questionnaire load together. Items in a construct must load together otherwise, that item or items may measure different ideas which you did not intend. On the other hand, CFA is working on confirming the constructs measured. I recommend SPSS for such analysis.