Okay, I could spend more time on this and reference Books Online, but I'm just going to give it a quick once over and list anything I have queries about.
Are extents always 64k? I thought this could be changed (not that I have ever tried to do so.)
Tempfile in Oracle is kinda equivalent to tempdb. Kinda.
Default tablespace in Oracle is kinda equivalent to [PRIMARY].
RAC: you can do similar sorts of things with federated servers.
SQL*Loader: bcp/SSIS. bcp is less capable than SQL*Loader, SSIS is far more capable.
Oracle Clusters: this is some stuff that's been hanging around in Oracle since at least version 6. I don't know of anyone who uses them. They just make life very complicated. It's a way of merging the storage of overlapping tables. When disc space was a premium, it made sense. These days, use data compression.
Validate/NoValidate: I think there's a SQL Server equivalent to this when you create a constraint to get it to check the data?
REF: why do we need to create hierarchical data stores in a relational database? Grrr. Anyway, you could possibly do some of this with a SQLVariant.
Dimension: Oracle has a chunk of data warehousing stuff built into the RDBMS. For SQL Server, you need to step outside into Analysis Services to get the equivalent.
Parameters: OUTPUT parameters in SQL Server are the equvalent to IN OUT in Oracle.
By default, indexes in both Oracle and SQL Server are B*Tree. You don't actually specify "B*Tree" in Oracle.
Compressed Indexes: is this available in SQL 2008? I know data compression is.
They haven't shown anything for SQL Server's INCLUDE() as part of an index.
Nothing about XML indexes (which I believe you can do on both platforms.)
For Reverse, they are referring to Oracle's reverse key indexes, where it reverses the order of the bytes in the key before storing them. This reduces index hot-spots and contention in high TP scenarios. Mostly useful for non-clustered indexes. Oracle and SQL Server have very different space allocation mechanisms, so this really isn't as much of an issue with SQL Server.
Libraries: there's still (I believe) the concept of a library (DLL) within SQL Server, as distinct from an assembly. Not that I would recommend using them!
AUTHID: I believe this has some similarity to execute as.
Check constraints are deprecated in SQL Server? Who knew!
Identity vs Sequence: oh yes, we could have some arguments over this one! Having used both, I prefer identity. It's easy enough to build a table to duplicate Sequences if you really need them.
Public Synonyms: this just creates a synonym in the "public" schema. In SQL Server, just create your synonym then grant rights to public.
Compressed Table: some new stuff with SQL 2008 here (the article IS comparing to SQL 2005).
Function Based Index: it says the equivalent in SQL Server is Computed Column, but that's not quite right. You need to create an index over the computed column to get the same effect.
External Table: I'd say OPENROWSET() is an equivalent.
Global Temporary Table (on commit delete rows): #table is mostly equivalent.
Global Temporary Table (on commit preserve rows): ##table is mostly equivalent.
XML Table: in SQL Server, equivalent would be a table with an XML column.
PCTUSED, PCTFREE, INITRANS, MAXTRANS: heh. Hangovers from an older Oracle. Not even Oracle uses these (much) any more.
Types: Oracle types are far more capable than SQL Server types, unless you are willing to step outside and into .net assembly-based types. In which case, you can have all of the equivalents to Oracle's types.
Varray, a table variable is similar.
SQL Server has ! ? Who knew...
Date operators: er, dateadd(), datediff()? They're not operators, they are functions. You can use + and - with dates too (if you specifically want operators).
Should I keep going?